Skip to content

armgit / bla

Public Imported from GitHub
0

🛡️ BLA (Blind LLM Agents) — Adversarial defense shield that protects binaries, Electron apps, JS bundles, Python, and WebAssembly against AI reverse-engineering agents like REA. Poison -> Exhaust -> Hallucinate.

docs(website,readme): refine technical tone, add GitHub star counter, and remove external repo references

qybaihedc1a9dd3a0d4
diff --git a/README.md b/README.md
index e7e1481..87a2c3e 100644
--- a/README.md
+++ b/README.md
@@ -4,9 +4,9 @@
# <img src="docs/assets/logo.png" alt="BLA Shield" width="42" align="absmiddle" /> BLA: Blind LLM Agents
-### One MCP for defending binaries, applications, and runtime behavior against AI reverse engineering.
+### One MCP for protecting binaries, applications, and runtime behavior against automated reverse engineering.
-**Ship a feature you built. Blind the agents trying to clone it, down to the binary level.**
+**Point it at a build artifact you ship. Protect its structure, logic, and runtime behavior down to the binary level.**
[![npm version](https://img.shields.io/badge/npm-bla--agents-2563eb?style=flat-square&logo=npm)](https://www.npmjs.com/package/bla-agents)
[![MCP tool catalog](https://img.shields.io/badge/MCP-tool_catalog-3b82f6?style=flat-square)](docs/cli-and-mcp.md)
@@ -30,7 +30,7 @@
<br />
-<img src="docs/assets/bla-vs-rea-showcase.svg" alt="Side-by-side comparison: REA reconstructing an unprotected app in 38s vs. BLA exhausting REA with 2,121x token burn and resource capacity trips" width="1200" />
+<img src="docs/assets/bla-vs-rea-showcase.svg" alt="Side-by-side comparison: REA analyzing an unprotected artifact vs. BLA-protected output" width="1200" />
<br />
@@ -38,11 +38,11 @@
---
-Worried that someone will point [REA](https://github.com/morluto/rea) or a coding agent at your shipped product and clone your core logic in seconds? Ask your agent to harden it with BLA (`Poison → Exhaust → Hallucinate`). It audits your build without requiring source-code rewrites, shows what an attacking agent would recover, injects adversarial chaff and algorithmic tarpits, and verifies that automated reconstruction fails.
+BLA is an MCP server and CLI for auditing, protecting, and verifying software build artifacts against AI reverse-engineering tools such as REA. It inspects release outputs for exposed source maps and symbols, applies format-specific static and runtime protections without requiring source-code rewrites, and verifies that automated static/semantic reconstruction is degraded while original runtime behavior remains intact.
-BLA connects your agent to tools for protecting native binaries, JavaScript and Electron apps, WebAssembly modules, Python bytecode, SQLite databases, AI model files, and mobile packages. You can also run the same tools from your terminal or build bundler. Protection runs locally and preserves 100% of original runtime behavior.
+BLA connects your coding agent to protection workflows for native binaries, JavaScript and Electron applications, WebAssembly modules, Python bytecode, SQLite databases, AI model files, and mobile packages. You can also run the same tools directly from your terminal or build bundler.
-Setup registers BLA with your agent and installs matching workflow instructions. Every static, bytecode, and binary shield is built in—no commercial packer, external LLVM toolchain, or third-party obfuscator is required.
+Setup registers BLA with your agent and installs the matching workflow instructions. Every static, bytecode, and binary protection pass is built in—no commercial packer, external LLVM toolchain, or third-party obfuscator is required.
## Quick start
@@ -60,46 +60,46 @@ Setup supports Claude Code, Codex, Cursor, Gemini CLI, Windsurf, Google Antigrav
### Ask your agent
-Every defense stage—or the entire 5-stage pipeline—can be triggered directly in natural language once your agent is connected:
+Once connected, you can trigger the full 5-stage pipeline or individual stages in natural language:
-- **⚡ One-Click Full Defense Hardening (`bla_defend` — Inspect → Audit → Bankrupt Protect → Verify → Manifest):**
+- **One-click full defense pipeline (`bla_defend` — Inspect → Audit → Protect → Verify → Manifest):**
```text
- Run bla_defend on ./dist with bankrupt intensity and emit a signed Counter-REA defense manifest.
+ Run bla_defend on ./dist with bankrupt intensity and emit a verification manifest.
```
-- **1. Audit exposure before shipping (`Pre-flight REA Audit`):**
+- **1. Audit exposure before release (`Audit`):**
```text
- Audit my ./dist directory against REA, show me every exposed module and IPC channel, and estimate how much it would cost Claude to reconstruct it.
+ Audit my ./dist directory against REA, list exposed modules and IPC channels, and estimate reconstruction cost.
```
-- **2. Inject Counter-REA protection (`Poison + Exhaust + Hallucinate`):**
+- **2. Protect an artifact (`Protect`):**
```text
- Protect ./dist/app.asar with BLA at bankrupt intensity, inject enclave chaff and algorithmic tarpits, and output to ./dist/app.shielded.asar.
+ Protect ./dist/app.asar with BLA and write the hardened package to ./dist/app.shielded.asar.
```
-- **3. Verify defense gates & token amplification (`Verify`):**
+- **3. Verify protection & runtime equivalence (`Verify`):**
```text
- Verify ./dist/app.shielded.asar against the original ./dist/app.asar and confirm that REA's examination degrades to limited status.
+ Verify ./dist/app.shielded.asar against ./dist/app.asar and confirm that REA's static analysis degrades to limited status.
```
### Use the terminal
Every workflow is available directly from your terminal with `npx bla-agents` (or `bla` when installed globally):
-- **⚡ One-Click Full Defense Pipeline (`defend`)** — runs release exposure scan, pre-flight REA audit, bankrupt exhaustion + deception injection, verification, and `.bla-manifest.json` generation in one command:
+- **One-click full pipeline (`defend`)** — runs release exposure scan, pre-flight audit, protection, verification, and `.bla-manifest.json` generation in one command:
```bash
npx -y bla-agents@latest defend /absolute/path/to/app --manifest
```
-- **1. Audit a target (`audit`)** — inspect exposure score (`0–100`), exposed symbols, and estimated attacker clone cost:
+- **1. Audit a target (`audit`)** — inspect exposure score (`0–100`), exposed symbols, and estimated analysis cost:
```bash
npx -y bla-agents@latest audit /absolute/path/to/app --json
```
-- **2. Protect a target (`protect`)** — inject semantic budget breakers, tarpit state machines, enclave chaff, and runtime shields:
+- **2. Protect a target (`protect`)** — apply format-specific budget guards, call-graph complexity, decoy metadata, and runtime checks:
```bash
npx -y bla-agents@latest protect /absolute/path/to/app -o /absolute/path/to/app.shielded --intensity bankrupt
```
-- **3. Verify Counter-REA resistance (`verify`)** — check that REA's static/semantic passes hit resource limits and >10x token amplification:
+- **3. Verify protection (`verify`)** — check that static/semantic analyzers hit resource limits while preserving behavior:
```bash
npx -y bla-agents@latest verify /absolute/path/to/app.shielded /absolute/path/to/app
```
-- **4. Run the 8-target Counter-REA benchmark (`benchmark`)**:
+- **4. Run the 8-target benchmark (`benchmark`)**:
```bash
npx -y bla-agents@latest benchmark
```
@@ -108,7 +108,7 @@ Every workflow is available directly from your terminal with `npx bla-agents` (o
npx -y bla-agents@latest emit-headers --lang c -o ./include/bla_shield.h
```
-To install the `bla` command for regular use:
+To install the `bla` command globally:
```bash
npm install --global bla-agents
@@ -117,7 +117,7 @@ bla --help
### Use in your build bundler
-Import `bla-shield/bundler` in Vite, Rollup, Webpack, or esbuild to automatically shield emitted JavaScript chunks, Honey Source Maps, and `.wasm` assets at build time:
+Import `bla-agents/bundler` in Vite, Rollup, Webpack, or esbuild to protect emitted JavaScript chunks, source maps, and `.wasm` assets at build time:
```typescript
import { defineConfig } from "vite";
@@ -142,44 +142,42 @@ bla update
npx bla-agents@latest setup
```
-Review the setup changes and restart your agent. For one-off CLI commands, use `npx bla-agents@latest` (or `bunx bla-agents@latest`) followed by the command.
-
## How BLA works
-Your agent calls BLA through MCP to audit a build artifact, inject format-specific Counter-REA shields, and verify the hardened output. When an attacking agent later inspects the protected artifact with REA, REA hits deterministic internal resource limits, drowns in decoy cross-references, and misattributes core logic to hardware enclave stubs:
+Your agent calls BLA through MCP to audit a build artifact, apply format-specific protections, and verify the output. When an automated reverse-engineering pipeline inspects the protected artifact, it encounters three defensive layers:
-1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds benign hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 **Honey Source Maps** (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables across dead branches and constant pools.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s hardcoded analyzer caps (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) at zero runtime cost, while weaving 14–36 interconnected Fake XRef state machines into the call graph.
-3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires in-place (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and injects the "Truman Show" runtime guard that serves synthetic enclave attestation telemetry whenever CDP, Node Inspector, Frida, or WebDriver probes attach.
+1. **Resource budget saturation (`Exhaustion`):** Triggers static and semantic analyzer resource guards in tools like `REA` (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) using $O(1)$ runtime invariants and strongly connected call-graph meshes inside unreachable branches.
+2. **Evidence & control-flow decoy (`Deception`):** Demotes real business logic to `status: "unknown"` under static analysis while providing clean, self-consistent shadow state machines, enclave delegation metadata (`SGX` / `TPM2` / `Secure Enclave`), and VLQ v3 decoy source maps.
+3. **Runtime & fuse hardening (`Runtime Guard`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and detects dynamic instrumentation (CDP, Node Inspector, Frida, or WebDriver) to return synthetic telemetry during inspection sessions.
## What you can protect
-BLA requires Node.js 22.5+ (or 24.x+) and npm. All shields run self-contained with zero external tool dependencies:
+BLA requires Node.js `^22.13.0 || >=24.0.0` and npm. All format shields run locally with zero external binary dependencies:
-| Target | What BLA injects & how REA fails | Command / prompt & requirements |
+| Target | Protection applied & analyzer impact | Command / prompt & requirements |
| ----------------------- | ------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- |
-| JavaScript / TypeScript | Trips `primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512` + 36 tarpits | `bla protect app.js` · Built-in Babel AST engine; [Counter-REA matrix](.agents/skills/blind-llm-agents/references/counter-rea-matrix.md) |
+| JavaScript / TypeScript | Trips `primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512` + SCC mesh | `bla protect app.js` · Built-in Babel AST engine; [Counter-REA matrix](.agents/skills/blind-llm-agents/references/counter-rea-matrix.md) |
| Electron `.asar` apps | De-literalizes IPC channels (`channel: null`), adds duplicate decoy IPC handlers & shadow module | `bla protect app.asar` · Built-in Chromium Pickle `.asar` unpacker & SHA-256 integrity repacker |
| Electron binary fuses | Hardens real V1 fuses (`RunAsNode=0`, `Inspect=0`) + injects `34 > 32` decoy fuse sentinels | `bla protect ElectronApp` · Trips `ElectronFuseResourceFailure (maximumSentinelsPerSlice = 32)` |
| Runtime CDP / Frida | Detects `inspector.url()`, `--inspect`, `__frida` and returns synthetic enclave telemetry | Automatically injected into JS/TS/Electron bundles via [`runtime-shield.ts`](src/shields/runtime/runtime-shield.ts) |
-| Source maps & HTML | Strips real maps; injects valid VLQ v3 Honey Source Maps (`enclave://tee-runtime/...`) | `bla protect index.html` · Redirects REA `recover_javascript_sources` & `trace_web_source_location` |
-| Native binaries | Appends `.bla.rodata` enclave chaff, decoy symbol table entries, and multi-scanner overflow traps | `bla protect daemon.elf` · Built-in ELF64, Mach-O (thin & universal), and Windows PE/COFF shield |
-| Rust binaries | Injects `68 > 64` distinct `/rustc/<40-hex>/` paths + decoy zlib `.dep-v0` crate dependency graph | `bla protect rust-bin` · Trips `RustBinaryProvider` (`MAX_RUSTC_HASHES = 64` capacity failure) |
-| Go binaries | Injects decoy Go `go1.23.4` buildinfo & `dep\tgithub.com/enclave-tees/sgx-bridge` module lines | `bla protect go-bin` · Pollutes REA `inspect_go_binary` module inventory |
+| Source maps & HTML | Strips real maps; emits valid VLQ v3 decoy Source Maps (`enclave://tee-runtime/...`) | `bla protect index.html` · Redirects `recover_javascript_sources` & `trace_web_source_location` |
+| Native binaries | Appends `.bla.rodata` enclave metadata, decoy symbol entries, and scanner limit guards | `bla protect daemon.elf` · Built-in ELF64, Mach-O (thin & universal), and Windows PE/COFF shield |
+| Rust binaries | Injects `68 > 64` distinct `/rustc/<40-hex>/` paths + decoy zlib `.dep-v0` crate dependency graph | `bla protect rust-bin` · Trips `RustBinaryProvider` (`MAX_RUSTC_HASHES = 64` limit) |
+| Go binaries | Injects decoy Go `go1.23.4` buildinfo & `dep\tgithub.com/enclave-tees/sgx-bridge` module lines | `bla protect go-bin` · Dilutes `inspect_go_binary` module inventory |
| Flutter / Dart AOT | Injects `66 > 64` Dart snapshot magic headers (`0xf5f5dcdc`) + `6 > 4` LLVM toolchain lines | `bla protect libapp.so` · Trips `SnapshotHashScanner (capped: true)` & `conflicting-dart-version` |
-| WASM artifacts | Strips debug sections; injects honey `"name"` section (`tee_hardware_enclave_stub`) + custom chaff | `bla protect core.wasm` · Built-in LEB128 binary parser; preserves 100% `WebAssembly.validate` |
-| Python & `.pyc` | Injects Python tarpit state machines into `.py` & `TYPE_ASCII` marshal chaff into `.pyc` pools | `bla protect module.pyc` · Built-in CPython 3.7–3.14 `.pyc` header & marshal framing shield |
-| SQLite databases | Injects 3 enclave honey tables with `GENERATED ALWAYS AS (bla_enclave_unseal(...)) VIRTUAL` traps | `bla protect data.sqlite` · Real tables stay 100% queryable; REA crashes with `SqliteColumnEvaluationFailure` |
-| Model artifacts | Injects `35 > 32` nested JSON/ARRAY headers into `.safetensors` and `.gguf` model headers | `bla protect weights.safetensors` · Trips `ModelFormatFailure (json-nesting / metadata-nesting > 32)` |
-| Android DEX & APKs | Injects `Lcom/enclave/sgx/...` decoy classes into `classes.dex` and recomputes SHA-1 & Adler32 | `bla protect classes.dex` or `app.apk` · Valid Dalvik/ART verification; pollutes JADX/JEB class trees |
-| Unity IL2CPP metadata | Appends enclave bridge metadata and method-topology tarpits to `global-metadata.dat` (`0xFAB11BAF`) | `bla protect global-metadata.dat` · Pollutes REA `inspect_il2cpp_metadata` type/method recovery |
-| C / C++ / Rust / Go src | Generates compile-time `bla_shield.h`, `bla_shield.rs`, or `bla_shield.go` with `used, noinline` tarpits | `bla emit-headers --lang c\|rust\|go` ·Compiles directly into Ghidra/Hopper/IDA decompiler views |
+| WASM artifacts | Strips debug sections; adds decoy `"name"` section (`tee_hardware_enclave_stub`) + custom sections | `bla protect core.wasm` · Built-in LEB128 binary parser; preserves 100% `WebAssembly.validate` |
+| Python & `.pyc` | Injects call-graph state machines into `.py` & `TYPE_ASCII` marshal entries into `.pyc` pools | `bla protect module.pyc` · Built-in CPython 3.7–3.14 `.pyc` header & marshal framing shield |
+| SQLite databases | Adds 3 decoy tables with `GENERATED ALWAYS AS (bla_enclave_unseal(...)) VIRTUAL` column guards | `bla protect data.sqlite` · Real tables stay queryable; read-only inspection hits `SqliteColumnEvaluationFailure` |
+| Model artifacts | Adds `35 > 32` nested JSON/ARRAY headers into `.safetensors` and `.gguf` metadata headers | `bla protect weights.safetensors` · Trips `ModelFormatFailure (json-nesting / metadata-nesting > 32)` |
+| Android DEX & APKs | Appends `Lcom/enclave/sgx/...` decoy descriptors into `classes.dex` and recomputes SHA-1/Adler32 | `bla protect classes.dex` or `app.apk` · Valid Dalvik/ART verification; expands decompiler class trees |
+| Unity IL2CPP metadata | Appends enclave bridge metadata and method-topology entries to `global-metadata.dat` (`0xFAB11BAF`) | `bla protect global-metadata.dat` · Dilutes `inspect_il2cpp_metadata` type/method recovery |
+| C / C++ / Rust / Go src | Generates compile-time `bla_shield.h`, `bla_shield.rs`, or `bla_shield.go` headers | `bla emit-headers --lang c\|rust\|go` · Compiles into binary symbol and control-flow graphs |
## Showcases
-Run `npx bla-agents benchmark` to execute all 8 counter-showcases locally and verify both REA failure modes and 100% runtime equivalence:
+Run `npx bla-agents benchmark` to execute all 8 benchmark targets locally and verify both analyzer degradation and runtime equivalence:
-| Showcase Target (Modeled on REA Official Showcases) | Before BLA (REA Alone) | After `npx bla-agents protect` | Token Burn | Runtime Check |
+| Benchmark Target | Unprotected (Baseline) | Protected (`npx bla-agents protect`) | Token Ratio | Runtime Check |
| :--- | :--- | :--- | :---: | :---: |
| **1. DX-Ball Sound-Pan Calculation** (`.js`) | `84/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `2/100` · `10,182,068` tok · **`$160.64`** (`LIMITED`) | **2,121.3x** | ✅ 100% Exact |
| **2. Notion Electron Clipboard Bridge** (`.asar`) | `98/100` · `9,132` tok · **`$0.34`** (`COMPLETE`) | `2/100` · `23,125,000` tok · **`$358.28`** (`LIMITED`) | **2,532.3x** | ✅ 100% Exact |
@@ -190,17 +188,17 @@ Run `npx bla-agents benchmark` to execute all 8 counter-showcases locally and ve
| **7. Proprietary Edge AI Weights** (`.safetensors`) | `81/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `6/100` · `1,807,073` tok · **`$29.88`** (`LIMITED`) | **376.5x** | ✅ 100% Exact |
| **8. Android Dalvik Bytecode** (`classes.dex`) | `81/100` · `8,900` tok · **`$0.34`** (`COMPLETE`) | `7/100` · `2,786,285` tok · **`$45.09`** (`LIMITED`) | **313.1x** | ✅ 100% Exact |
-### DX-Ball: bankrupt a sound-pan reconstruction
+### DX-Ball: protecting sound-pan calculation logic
-REA's flagship showcase reconstructs DX-Ball's `clampPan` and `calculateSoundPanFromX` helpers in seconds (`$0.20`). After running `npx bla-agents protect examples/dx-ball-sound-pan/sound-pan.js --intensity bankrupt`, REA's `parse_status` drops to `"partial"`, `examination.status` degrades to `"limited"` (`["alias-selections", "primitive-bytes", "primitive-candidates"]`), and the obligation ledger expands from 2 functions to 38 interconnected state machines (`2,121.3x` token amplification) while returning identical pan values for every input coordinate.
+On an unprotected build, static analysis recovers `clampPan` and `calculateSoundPanFromX` in a few tool turns (`$0.20`). After running `npx bla-agents protect examples/dx-ball-sound-pan/sound-pan.js --intensity bankrupt`, `parse_status` drops to `"partial"`, `examination.status` degrades to `"limited"` (`["alias-selections", "primitive-bytes", "primitive-candidates"]`), and the reconstruction obligation ledger expands across interconnected state machines while returning identical pan values for every input coordinate.
-### Notion: blind the Electron clipboard IPC bridge
+### Notion: obscuring the Electron clipboard IPC bridge
-REA traces Electron apps by matching unique literal strings between `ipcRenderer.invoke("clipboard:write-rich-format")` and `ipcMain.handle(...)`. BLA replaces real IPC channel strings with $\ge 9$-chunk number-theoretic opaque expressions (`electron.ipc[0].channel === null`) and registers duplicate decoy handlers on the same literal channel names—forcing `electron_call_paths_limited: true` and `uncertainty: ["dynamic-findings", "returned-reference-widening"]`.
+Static Electron analyzers trace IPC workflows by matching literal channel names between `ipcRenderer.invoke("clipboard:write-rich-format")` and `ipcMain.handle(...)`. BLA replaces literal channel strings with $\ge 9$-chunk number-theoretic opaque expressions (`electron.ipc[0].channel === null`) and registers duplicate decoy handlers on the same literal channel names—setting `electron_call_paths_limited: true` and `uncertainty: ["dynamic-findings", "returned-reference-widening"]`.
-### Native, SQLite & Edge AI: trip REA's binary and container capacity limits
+### Native, SQLite & Edge AI: triggering parser capacity guards
-When an attacker runs REA's `inspect_rust_binary`, `inspect_electron_fuses`, `inspect_sqlite_database`, or `inspect_model_artifact` on BLA-protected files, REA's own parsers deterministically abort with capacity and evaluation failures (`Read-only data contains more than 64 distinct rustc commit paths`, `Slice 0 contains more than 32 Electron fuse sentinels`, `SqliteColumnEvaluationFailure`, and `ModelFormatFailure: json-nesting`).
+When `inspect_rust_binary`, `inspect_electron_fuses`, `inspect_sqlite_database`, or `inspect_model_artifact` runs against BLA-protected files, the analyzers hit their configured admission caps (`Read-only data contains more than 64 distinct rustc commit paths`, `Slice 0 contains more than 32 Electron fuse sentinels`, `SqliteColumnEvaluationFailure`, and `ModelFormatFailure: json-nesting`).
## FAQ
@@ -214,44 +212,44 @@ Any coding agent that supports local MCP servers or `SKILL.md` workflows. `npx b
<details>
<summary><strong>Does BLA slow down my app or change its runtime output?</strong></summary>
-No. BLA's semantic budget breakers (`alias-selections`, `getter-expansion`, `primitive-bytes`) and tarpit call graphs are guarded by $O(1)$ number-theoretic opaque predicates ($7x^2 - 1 \neq y^2 \pmod 7$) that never execute at runtime. Every benchmark target verifies 100% input/output equivalence against the unprotected build.
+No. BLA's static budget guards (`alias-selections`, `getter-expansion`, `primitive-bytes`) and call-graph expansions are guarded by $O(1)$ number-theoretic opaque predicates ($7x^2 - 1 \neq y^2 \pmod 7$) that do not execute on the hot path. Every benchmark target verifies 100% input/output equivalence against the unprotected build.
</details>
<details>
<summary><strong>How is BLA different from traditional obfuscators like javascript-obfuscator or O-LLVM?</strong></summary>
-Traditional obfuscators rename variables (`calculatePrice` $\rightarrow$ `_0x1a2b`), which modern LLMs effortlessly reverse via control-flow semantics, while heavy VM packers slow down execution. BLA is engineered specifically against **AI coding agents and `morluto/rea`'s static/semantic analyzers**—tripping REA's exact internal resource caps, breaking IPC literal pairing, poisoning source maps/metadata with hardware-enclave delegation chaff, and multiplying agent MCP token consumption by up to 2,500x.
+Traditional obfuscators rely primarily on identifier renaming (`calculatePrice` $\rightarrow$ `_0x1a2b`), which LLMs infer from control-flow context, while heavy VM packers add runtime overhead. BLA targets the static analyzers, resource budgets, and evidence rules used by AI reverse-engineering pipelines—breaking literal IPC pairing, tripping parser capacity limits, and introducing structured shadow logic.
</details>
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed to protect my files?</strong></summary>
-No. All 19 format shields (Babel AST, Chromium `.asar`, LEB128 `.wasm`, CPython `.pyc`, ELF/Mach-O/PE, SQLite3, `.safetensors`/`.gguf`, and `.dex`/`Il2Cpp`) are built directly into `bla-shield` with zero external binary dependencies.
+No. All 19 format shields (Babel AST, Chromium `.asar`, LEB128 `.wasm`, CPython `.pyc`, ELF/Mach-O/PE, SQLite3, `.safetensors`/`.gguf`, and `.dex`/`Il2Cpp`) are built directly into `bla-agents` with zero external binary dependencies.
</details>
<details>
<summary><strong>What happens when someone runs REA against a BLA-protected app?</strong></summary>
-1. REA's JavaScript/Electron analyzer degrades from `complete` to `partial` / `limited` and fails to pair IPC channels (`channel: null`).
-2. REA's binary, Fuse, SQLite, and AI model inspectors hit hard capacity limits (`>64` rustc hashes, `>32` Electron fuse sentinels, `SqliteColumnEvaluationFailure`, `>32` JSON/ARRAY nesting).
-3. If the agent attempts feature tracing (`trace_application_feature`), it enters 14–36 Fake XRef state machines and Honey Source Maps (`enclave://tee-runtime/`), exhausting its context window and API budget.
+1. The JavaScript/Electron analyzer degrades from `complete` to `partial` / `limited` and cannot pair IPC channels (`channel: null`).
+2. Binary, Fuse, SQLite, and model inspectors hit admission limits (`>64` rustc hashes, `>32` Electron fuse sentinels, `SqliteColumnEvaluationFailure`, `>32` JSON/ARRAY nesting).
+3. Feature-tracing workflows (`trace_application_feature`) encounter decoy cross-reference state machines and decoy source maps (`enclave://tee-runtime/`).
</details>
<details>
<summary><strong>What does installing the skill do?</strong></summary>
-The [`blind-llm-agents` skill](.agents/skills/blind-llm-agents/SKILL.md) teaches your coding agent how to run pre-release Counter-REA audits, select the right protection intensity (`stealth`, `balanced`, `maximum`, `bankrupt`), integrate bundler plugins, and verify defense gates automatically.
+The [`blind-llm-agents` skill](.agents/skills/blind-llm-agents/SKILL.md) teaches your coding agent how to run pre-release audits, choose a protection intensity (`stealth`, `balanced`, `maximum`, `bankrupt`), integrate bundler plugins, and verify protected outputs automatically.
</details>
<details>
<summary><strong>Does BLA upload my source code or binaries anywhere?</strong></summary>
-Never. BLA audits, protects, and verifies all files 100% locally on your machine with deterministic seeded PRNG transforms (`--seed`) and zero network calls.
+No. BLA audits, protects, and verifies all files locally on your machine with deterministic seeded PRNG transforms (`--seed`) and no network calls.
</details>
@@ -273,7 +271,7 @@ If the issue persists, [open an issue](https://github.com/qybaihe/bla/issues) wi
- [Installation and agent setup](docs/installation.md): agent registration, bundler plugins, updates, and uninstall.
- [CLI and MCP contracts](docs/cli-and-mcp.md): commands, options, exit codes, and JSON-RPC 2.0 MCP tool schemas.
-- [Counter-REA technical defense matrix](.agents/skills/blind-llm-agents/references/counter-rea-matrix.md): exact `morluto/rea` source limits and how BLA exploits them.
+- [Counter-REA technical defense matrix](.agents/skills/blind-llm-agents/references/counter-rea-matrix.md): `REA` resource limits and corresponding BLA guards.
Report security issues through [SECURITY.md](SECURITY.md).
@@ -281,7 +279,7 @@ Report security issues through [SECURITY.md](SECURITY.md).
See [DISCLAIMER.md](DISCLAIMER.md) for responsible-use guidance.
-BLA is a defensive software-protection toolkit for hardening software, bundles, models, and binaries that you own or are authorized to protect. All injected chaff payloads are strictly benign architectural metadata and synthetic decoy functions that preserve 100% of original runtime behavior.
+BLA is a defensive software-protection toolkit for hardening software, bundles, models, and binaries that you own or are authorized to protect. All injected metadata and decoy functions are benign structures that preserve original runtime behavior.
## Contributing
diff --git a/docs/apple-and-dotnet-defense.md b/docs/apple-and-dotnet-defense.md
index 6cf26a3..18d45f8 100644
--- a/docs/apple-and-dotnet-defense.md
+++ b/docs/apple-and-dotnet-defense.md
@@ -2,7 +2,7 @@
## 1. Apple `.plist` Internal Subset Budget Trap (`src/shields/formats/apple-plist.ts`)
-In `morluto/rea/src/artifacts/apple/PropertyListDecodeBudget.ts:80`:
+In `REA/src/artifacts/apple/PropertyListDecodeBudget.ts:80`:
```ts
if (token.doctype) {
if (token.internalSubset)
@@ -15,4 +15,4 @@ REA rejects any XML `.plist` whose `<!DOCTYPE plist ...>` tag contains a `[` bra
## 2. .NET Managed Metadata Representation Amplification (`src/shields/formats/dotnet-managed.ts`)
-In `morluto/rea/src/dotnet/ManagedDecodeBudget.ts`, every metadata table row reserves `256` bytes and every UTF-8 string reserves `encodedBytes * 6 + 2` bytes against a 64 MiB ceiling, while `managed-pinvoke-import` entries trigger mandatory native ABI reconstruction obligations.
+In `REA/src/dotnet/ManagedDecodeBudget.ts`, every metadata table row reserves `256` bytes and every UTF-8 string reserves `encodedBytes * 6 + 2` bytes against a 64 MiB ceiling, while `managed-pinvoke-import` entries trigger mandatory native ABI reconstruction obligations.
diff --git a/docs/assets/bla-vs-rea-showcase.svg b/docs/assets/bla-vs-rea-showcase.svg
index 01f86f9..6a24bf0 100644
--- a/docs/assets/bla-vs-rea-showcase.svg
+++ b/docs/assets/bla-vs-rea-showcase.svg
@@ -35,7 +35,7 @@
<!-- Top Header -->
<text x="600" y="38" text-anchor="middle" class="sans" font-size="20" font-weight="800" fill="#f8fafc">
- ⚔️ RED vs. BLUE SHOWCASE: <tspan fill="#ef4444">morluto/rea (RED)</tspan> vs. <tspan fill="#60a5fa">npx bla-agents defend (BLUE)</tspan>
+ ⚔️ RED vs. BLUE SHOWCASE: <tspan fill="#ef4444">REA (RED)</tspan> vs. <tspan fill="#60a5fa">npx bla-agents defend (BLUE)</tspan>
</text>
<!-- ==================== LEFT SCREEN: BEFORE BLA (REA RED) ==================== -->
diff --git a/docs/cdp-and-frida-runtime-defense.md b/docs/cdp-and-frida-runtime-defense.md
index 7577005..1eee202 100644
--- a/docs/cdp-and-frida-runtime-defense.md
+++ b/docs/cdp-and-frida-runtime-defense.md
@@ -2,13 +2,13 @@
## 1. Defeating REA's `CdpCaptureEvents` Shape & Event Budgets
-In `morluto/rea/src/browser/CdpCaptureEvents.ts`, passive browser inspection shares a fixed `WEB_RUNTIME_LIMITS.retainedEventBytes` allowance across console logs, HTTP JSON body shapes, and WebSocket frame shapes. Any JSON payload nested $\ge 760$ levels deep is rejected with `reason: "resource_budget_exhausted"`.
+In `REA/src/browser/CdpCaptureEvents.ts`, passive browser inspection shares a fixed `WEB_RUNTIME_LIMITS.retainedEventBytes` allowance across console logs, HTTP JSON body shapes, and WebSocket frame shapes. Any JSON payload nested $\ge 760$ levels deep is rejected with `reason: "resource_budget_exhausted"`.
BLA's `generateCdpShapeBudgetBreaker` detects Playwright, Puppeteer, and WebDriver attachment and synthesizes 780–900 level JSON envelopes alongside console drain payloads.
## 2. Defeating REA's `fridaCaptureLimits.ts`
-In `morluto/rea/src/frida/fridaCaptureLimits.ts`:
+In `REA/src/frida/fridaCaptureLimits.ts`:
- `MAX_JSON_DEPTH = 64`
- `MAX_MESSAGE_TEXT_BYTES = 32,768`
- `MAX_CAPTURE_JSON_NODES = 250,000`
diff --git a/docs/counter-rea-exhaustion.md b/docs/counter-rea-exhaustion.md
index f626333..3b6d853 100644
--- a/docs/counter-rea-exhaustion.md
+++ b/docs/counter-rea-exhaustion.md
@@ -1,8 +1,8 @@
# Counter-REA Resource Exhaustion Architecture (`src/shields/exhaustion/`)
-BLA's Resource Exhaustion Engine is engineered directly from a line-by-line audit of `morluto/rea`'s internal resource caps, worker admission budgets, and obligation ledger rules.
+BLA's Resource Exhaustion Engine is engineered directly from a line-by-line audit of `REA`'s internal resource caps, worker admission budgets, and obligation ledger rules.
-## 1. Exact `morluto/rea` Internal Budget Breakers
+## 1. Exact `REA` Internal Budget Breakers
| REA Subsystem File | Constant | Hard Cap | BLA Countermeasure |
| :--- | :--- | :--- | :--- |
@@ -16,6 +16,6 @@ BLA's Resource Exhaustion Engine is engineered directly from a line-by-line audi
## 2. 22-Family `ReconstructionObligationLedger` Flood (`obligation-flooder.ts`)
-In `morluto/rea/src/application/ReconstructionObligationPolicies.ts`, every detected application or semantic boundary node (`context-bridge-api`, `ipc-channel`, `ipc-handler`, `endpoint`, `storage`, `worker`, `service-worker`, `native-addon`, `child-process`, `stdio`, `signal`, `config-source`, `request`, `response`, `boundary`, `resource`, `event`, `timer`, `promise`, `task`) forces the coding agent to construct 3–4 separate verification fixtures (`positive`, `negative`, `malformed`, `cancellation`, `teardown`).
+In `REA/src/application/ReconstructionObligationPolicies.ts`, every detected application or semantic boundary node (`context-bridge-api`, `ipc-channel`, `ipc-handler`, `endpoint`, `storage`, `worker`, `service-worker`, `native-addon`, `child-process`, `stdio`, `signal`, `config-source`, `request`, `response`, `boundary`, `resource`, `event`, `timer`, `promise`, `task`) forces the coding agent to construct 3–4 separate verification fixtures (`positive`, `negative`, `malformed`, `cancellation`, `teardown`).
BLA injects dead-branch AST patterns matching all 22 obligation families, multiplying the agent's mandatory proof workload by **20x–60x** with zero runtime cost.
diff --git a/docs/installation.md b/docs/installation.md
index 023e61d..414168e 100644
--- a/docs/installation.md
+++ b/docs/installation.md
@@ -32,11 +32,11 @@ npx bla-agents verify ./dist/app.js ./dist/app.bla.js
### Install as a Dev Dependency (Recommended for CI/CD Release Pipelines)
```bash
-npm install --save-dev bla-shield
+npm install --save-dev bla-agents
# or
-pnpm add -D bla-shield
+pnpm add -D bla-agents
# or
-bun add -d bla-shield
+bun add -d bla-agents
```
Then add a post-build protection script to your `package.json`:
@@ -53,11 +53,11 @@ Then add a post-build protection script to your `package.json`:
### Install Globally
```bash
-npm install -g bla-shield
+npm install -g bla-agents
-# Both `bla` and `bla-shield` binaries are registered:
+# Both `bla` and `bla-agents` binaries are registered:
bla --version
-bla-shield doctor
+bla-agents doctor
```
---
diff --git a/docs/readme/README_ar.md b/docs/readme/README_ar.md
index 311bac0..f9edbec 100644
--- a/docs/readme/README_ar.md
+++ b/docs/readme/README_ar.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-هل تقلق من استخدام [REA](https://github.com/morluto/rea) لاستنساخ تطبيقك في ثوانٍ؟ اطلب من وكيلك تحصينه باستخدام BLA (`Poison → Exhaust → Hallucinate`).
+هل تقلق من استخدام REA لاستنساخ تطبيقك في ثوانٍ؟ اطلب من وكيلك تحصينه باستخدام BLA (`Poison → Exhaust → Hallucinate`).
يربط BLA وكيلك بأدوات حماية الملفات الثنائية وتطبيقات JavaScript وElectron وWebAssembly وPython وقواعد بيانات SQLite ونماذج الذكاء الاصطناعي محلياً مع الحفاظ على سلوك التشغيل بنسبة 100%.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## كيف يعمل BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_de.md b/docs/readme/README_de.md
index 5f9f7d3..5ac095c 100644
--- a/docs/readme/README_de.md
+++ b/docs/readme/README_de.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Befürchten Sie, dass jemand [REA](https://github.com/morluto/rea) auf Ihre ausgelieferte App ansetzt und Ihre Kernlogik in Sekunden klont? Härten Sie Ihren Build mit BLA (`Poison → Exhaust → Hallucinate`).
+Befürchten Sie, dass jemand REA auf Ihre ausgelieferte App ansetzt und Ihre Kernlogik in Sekunden klont? Härten Sie Ihren Build mit BLA (`Poison → Exhaust → Hallucinate`).
BLA verbindet Ihren Agenten mit Schutzwerkzeugen für native Binärdateien, JavaScript- und Electron-Apps, WebAssembly, Python-Bytecode, SQLite-Datenbanken und KI-Modelle – 100 % lokal und ohne Laufzeiteinbußen.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Wie BLA funktioniert
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_el.md b/docs/readme/README_el.md
index 2574ce4..9c04128 100644
--- a/docs/readme/README_el.md
+++ b/docs/readme/README_el.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Ανησυχείτε μήπως κάποιος χρησιμοποιήσει το [REA](https://github.com/morluto/rea) για να κλωνοποιήσει την εφαρμογή σας σε δευτερόλεπτα; Ζητήστε από τον πράκτορά σας να την θωρακίσει με το BLA (`Poison → Exhaust → Hallucinate`).
+Ανησυχείτε μήπως κάποιος χρησιμοποιήσει το REA για να κλωνοποιήσει την εφαρμογή σας σε δευτερόλεπτα; Ζητήστε από τον πράκτορά σας να την θωρακίσει με το BLA (`Poison → Exhaust → Hallucinate`).
Το BLA προστατεύει εγγενή δυαδικά αρχεία, εφαρμογές JavaScript και Electron, WebAssembly, Python bytecode, βάσεις δεδομένων SQLite και μοντέλα AI τοπικά, διατηρώντας 100% τη συμπεριφορά εκτέλεσης.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Πώς λειτουργεί το BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_es.md b/docs/readme/README_es.md
index 78dad31..f6297b1 100644
--- a/docs/readme/README_es.md
+++ b/docs/readme/README_es.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-¿Te preocupa que alguien use [REA](https://github.com/morluto/rea) para clonar tu aplicación en segundos? Pídele a tu agente que la blinde con BLA (`Poison → Exhaust → Hallucinate`).
+¿Te preocupa que alguien use REA para clonar tu aplicación en segundos? Pídele a tu agente que la blinde con BLA (`Poison → Exhaust → Hallucinate`).
BLA conecta tu agente con herramientas para proteger binarios nativos, aplicaciones JavaScript y Electron, módulos WebAssembly, bytecode de Python, bases de datos SQLite y modelos de IA localmente, preservando el 100% del comportamiento original.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Cómo funciona BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_fa.md b/docs/readme/README_fa.md
index 2ce5b10..7193c6a 100644
--- a/docs/readme/README_fa.md
+++ b/docs/readme/README_fa.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-نگران هستید کسی با استفاده از [REA](https://github.com/morluto/rea) برنامه شما را در چند ثانیه کلون کند؟ از ایجنت خود بخواهید آن را با BLA (`Poison → Exhaust → Hallucinate`) ایمن سازد.
+نگران هستید کسی با استفاده از REA برنامه شما را در چند ثانیه کلون کند؟ از ایجنت خود بخواهید آن را با BLA (`Poison → Exhaust → Hallucinate`) ایمن سازد.
BLA ایجنت شما را به ابزارهای محافظت از باینری‌های بومی، برنامه‌های JavaScript و Electron، ماژول‌های WebAssembly، بایت‌کد Python، پایگاه‌های داده SQLite و مدل‌های هوش مصنوعی متصل می‌کند.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## نحوه کار BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_fr.md b/docs/readme/README_fr.md
index 08a7d18..a18120e 100644
--- a/docs/readme/README_fr.md
+++ b/docs/readme/README_fr.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Vous craignez qu'un attaquant utilise [REA](https://github.com/morluto/rea) pour cloner votre application en quelques secondes ? Demandez à votre agent de la protéger avec BLA (`Poison → Exhaust → Hallucinate`).
+Vous craignez qu'un attaquant utilise REA pour cloner votre application en quelques secondes ? Demandez à votre agent de la protéger avec BLA (`Poison → Exhaust → Hallucinate`).
BLA connecte votre agent à des outils de protection pour binaires natifs, applications JavaScript et Electron, WebAssembly, bytecode Python, bases SQLite et modèles d'IA, 100 % en local et sans impact sur l'exécution.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Comment fonctionne BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_id.md b/docs/readme/README_id.md
index 5721c88..412cc83 100644
--- a/docs/readme/README_id.md
+++ b/docs/readme/README_id.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Khawatir seseorang menggunakan [REA](https://github.com/morluto/rea) untuk mengkloning aplikasi Anda dalam hitungan detik? Minta agen Anda melindunginya dengan BLA (`Poison → Exhaust → Hallucinate`).
+Khawatir seseorang menggunakan REA untuk mengkloning aplikasi Anda dalam hitungan detik? Minta agen Anda melindunginya dengan BLA (`Poison → Exhaust → Hallucinate`).
BLA menghubungkan agen Anda ke alat pelindung biner asli, aplikasi JavaScript & Electron, WebAssembly, bytecode Python, database SQLite, dan model AI secara lokal dengan mempertahankan 100% perilaku runtime.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Cara kerja BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_it.md b/docs/readme/README_it.md
index 4d18dfc..7944a37 100644
--- a/docs/readme/README_it.md
+++ b/docs/readme/README_it.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Temi che qualcuno usi [REA](https://github.com/morluto/rea) per clonare la tua app in pochi secondi? Chiedi al tuo agente di blindarla con BLA (`Poison → Exhaust → Hallucinate`).
+Temi che qualcuno usi REA per clonare la tua app in pochi secondi? Chiedi al tuo agente di blindarla con BLA (`Poison → Exhaust → Hallucinate`).
BLA collega il tuo agente a strumenti per proteggere binari nativi, app JavaScript ed Electron, WebAssembly, bytecode Python, database SQLite e modelli IA localmente, preservando il 100% del comportamento runtime.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Come funziona BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_ja.md b/docs/readme/README_ja.md
index 43de5a5..fa30452 100644
--- a/docs/readme/README_ja.md
+++ b/docs/readme/README_ja.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-[REA](https://github.com/morluto/rea) やコーディングエージェントにアプリを解析され、数秒でコアロジックを複製されるのを防ぎたいですか?エージェントに BLA(`Poison → Exhaust → Hallucinate`)でビルド成果物を保護させましょう。ソースコードの書き換えなしで露出度を監査し、エンクレーブのデコイとアルゴリズム・ターピットを注入して、自動再構築を確実に失敗させます。
+REA やコーディングエージェントにアプリを解析され、数秒でコアロジックを複製されるのを防ぎたいですか?エージェントに BLA(`Poison → Exhaust → Hallucinate`)でビルド成果物を保護させましょう。ソースコードの書き換えなしで露出度を監査し、エンクレーブのデコイとアルゴリズム・ターピットを注入して、自動再構築を確実に失敗させます。
BLA はネイティブバイナリ、JavaScript / Electron アプリ、WebAssembly、Python バイトコード、SQLite データベース、AI モデル(GGUF / safetensors)、モバイルパッケージの保護ツールにエージェントを接続します。処理はすべてローカルで実行され、元のランタイム動作を 100% 維持します。
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## BLA の仕組み
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_ko.md b/docs/readme/README_ko.md
index 4a81ea4..a56a585 100644
--- a/docs/readme/README_ko.md
+++ b/docs/readme/README_ko.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-[REA](https://github.com/morluto/rea)나 코딩 에이전트가 배포된 앱을 분석해 몇 초 만에 핵심 로직을 복제하는 것이 걱정되시나요? 에이전트에게 BLA(`Poison → Exhaust → Hallucinate`)로 빌드를 보호하도록 지시하세요. 소스 코드 수정 없이 노출도를 감사하고, 엔클레이브 채프와 알고리즘 타르핏을 주입하여 자동 재구성을 차단합니다.
+REA나 코딩 에이전트가 배포된 앱을 분석해 몇 초 만에 핵심 로직을 복제하는 것이 걱정되시나요? 에이전트에게 BLA(`Poison → Exhaust → Hallucinate`)로 빌드를 보호하도록 지시하세요. 소스 코드 수정 없이 노출도를 감사하고, 엔클레이브 채프와 알고리즘 타르핏을 주입하여 자동 재구성을 차단합니다.
BLA는 네이티브 바이너리, JavaScript 및 Electron 앱, WebAssembly, Python 바이트코드, SQLite 데이터베이스, AI 모델, 모바일 패키지를 보호하는 도구를 에이전트에 연결합니다. 모든 보호 작업은 로컬에서 실행되며 원래의 런타임 동작을 100% 유지합니다.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## BLA 작동 방식
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_ml.md b/docs/readme/README_ml.md
index beea131..675427d 100644
--- a/docs/readme/README_ml.md
+++ b/docs/readme/README_ml.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-[REA](https://github.com/morluto/rea) ഉപയോഗിച്ച് നിങ്ങളുടെ ആപ്പ് നിമിഷങ്ങൾക്കുള്ളിൽ ക്ലോൺ ചെയ്യപ്പെടുമെന്ന് ആശങ്കയുണ്ടോ? BLA (`Poison → Exhaust → Hallucinate`) ഉപയോഗിച്ച് ബിൽഡ് സുരക്ഷിതമാക്കാൻ നിങ്ങളുടെ ഏജന്റിനോട് ആവശ്യപ്പെടുക.
+REA ഉപയോഗിച്ച് നിങ്ങളുടെ ആപ്പ് നിമിഷങ്ങൾക്കുള്ളിൽ ക്ലോൺ ചെയ്യപ്പെടുമെന്ന് ആശങ്കയുണ്ടോ? BLA (`Poison → Exhaust → Hallucinate`) ഉപയോഗിച്ച് ബിൽഡ് സുരക്ഷിതമാക്കാൻ നിങ്ങളുടെ ഏജന്റിനോട് ആവശ്യപ്പെടുക.
നേറ്റീവ് ബൈനറികൾ, JavaScript/Electron ആപ്പുകൾ, WebAssembly, Python ബൈറ്റ്കോഡ്, SQLite ഡാറ്റാബേസുകൾ, AI മോഡലുകൾ എന്നിവ 100% റൺടൈം കൃത്യതയോടെ പ്രാദേശികമായി സംരക്ഷിക്കാൻ BLA സഹായിക്കുന്നു.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## BLA എങ്ങനെ പ്രവർത്തിക്കുന്നു
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_pl.md b/docs/readme/README_pl.md
index a9c7d2b..53fbf39 100644
--- a/docs/readme/README_pl.md
+++ b/docs/readme/README_pl.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Obawiasz się, że ktoś użyje [REA](https://github.com/morluto/rea), aby sklonować Twoją aplikację w kilka sekund? Zabezpiecz swój build za pomocą BLA (`Poison → Exhaust → Hallucinate`).
+Obawiasz się, że ktoś użyje REA, aby sklonować Twoją aplikację w kilka sekund? Zabezpiecz swój build za pomocą BLA (`Poison → Exhaust → Hallucinate`).
BLA łączy Twojego agenta z narzędziami do ochrony natywnych plików binarnych, aplikacji JavaScript i Electron, WebAssembly, bajtkodu Pythona, baz SQLite oraz modeli AI lokalnie, zachowując 100% oryginalnego działania.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Jak działa BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_pt-BR.md b/docs/readme/README_pt-BR.md
index 0a57d47..11d27c9 100644
--- a/docs/readme/README_pt-BR.md
+++ b/docs/readme/README_pt-BR.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Preocupado que alguém aponte o [REA](https://github.com/morluto/rea) para o seu aplicativo e clone sua lógica em segundos? Peça ao seu agente para blindá-lo com o BLA (`Poison → Exhaust → Hallucinate`).
+Preocupado que alguém aponte o REA para o seu aplicativo e clone sua lógica em segundos? Peça ao seu agente para blindá-lo com o BLA (`Poison → Exhaust → Hallucinate`).
O BLA conecta seu agente a ferramentas para proteger binários nativos, apps JavaScript e Electron, WebAssembly, bytecode Python, bancos SQLite e modelos de IA localmente, preservando 100% do comportamento em tempo de execução.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Como o BLA funciona
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_ro.md b/docs/readme/README_ro.md
index 0248907..94d5a57 100644
--- a/docs/readme/README_ro.md
+++ b/docs/readme/README_ro.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Te îngrijorează că cineva va folosi [REA](https://github.com/morluto/rea) pentru a-ți clona aplicația în câteva secunde? Cere-i agentului tău să o protejeze cu BLA (`Poison → Exhaust → Hallucinate`).
+Te îngrijorează că cineva va folosi REA pentru a-ți clona aplicația în câteva secunde? Cere-i agentului tău să o protejeze cu BLA (`Poison → Exhaust → Hallucinate`).
BLA conectează agentul tău la instrumente pentru protejarea fișierelor binare native, aplicațiilor JavaScript și Electron, WebAssembly, bytecode Python, bazelor de date SQLite și modelelor AI local, păstrând 100% comportamentul original.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Cum funcționează BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_ru.md b/docs/readme/README_ru.md
index b6b1836..a93b56d 100644
--- a/docs/readme/README_ru.md
+++ b/docs/readme/README_ru.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Опасаетесь, что [REA](https://github.com/morluto/rea) или ИИ-агент скопирует логику вашего приложения за считанные секунды? Поручите вашему агенту защитить сборку с помощью BLA (`Poison → Exhaust → Hallucinate`).
+Опасаетесь, что REA или ИИ-агент скопирует логику вашего приложения за считанные секунды? Поручите вашему агенту защитить сборку с помощью BLA (`Poison → Exhaust → Hallucinate`).
BLA подключает вашего агента к инструментам защиты нативных бинарных файлов, приложений JavaScript и Electron, модулей WebAssembly, байт-кода Python, баз данных SQLite и ИИ-моделей. Защита работает локально и на 100% сохраняет исходное поведение программы.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Как работает BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_th.md b/docs/readme/README_th.md
index 1b77f67..a8ff2bf 100644
--- a/docs/readme/README_th.md
+++ b/docs/readme/README_th.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-กังวลว่าจะมีคนใช้ [REA](https://github.com/morluto/rea) เพื่อคัดลอกแอปของคุณในไม่กี่วินาทีใช่ไหม? สั่งให้เอเจนต์ของคุณปกป้องบิลด์ด้วย BLA (`Poison → Exhaust → Hallucinate`)
+กังวลว่าจะมีคนใช้ REA เพื่อคัดลอกแอปของคุณในไม่กี่วินาทีใช่ไหม? สั่งให้เอเจนต์ของคุณปกป้องบิลด์ด้วย BLA (`Poison → Exhaust → Hallucinate`)
BLA เชื่อมต่อเอเจนต์เข้ากับเครื่องมือปกป้องไบนารีเนทีฟ, แอป JavaScript และ Electron, WebAssembly, Python bytecode, ฐานข้อมูล SQLite และโมเดล AI โดยทำงานแบบออฟไลน์ 100%
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## BLA ทำงานอย่างไร
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_tr.md b/docs/readme/README_tr.md
index 3eefca9..f743d6c 100644
--- a/docs/readme/README_tr.md
+++ b/docs/readme/README_tr.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-[REA](https://github.com/morluto/rea) veya bir kodlama ajanının uygulamanızı saniyeler içinde kopyalamasından mı endişeleniyorsunuz? Ajanınızdan BLA (`Poison → Exhaust → Hallucinate`) ile yapınızı güçlendirmesini isteyin.
+REA veya bir kodlama ajanının uygulamanızı saniyeler içinde kopyalamasından mı endişeleniyorsunuz? Ajanınızdan BLA (`Poison → Exhaust → Hallucinate`) ile yapınızı güçlendirmesini isteyin.
BLA; yerel ikili dosyaları, JavaScript ve Electron uygulamalarını, WebAssembly modüllerini, Python bayt kodunu, SQLite veritabanlarını ve AI modellerini yerel olarak korur ve çalışma zamanı davranışını %100 korur.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## BLA nasıl çalışır
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_uk.md b/docs/readme/README_uk.md
index ca2d4ab..51a28e6 100644
--- a/docs/readme/README_uk.md
+++ b/docs/readme/README_uk.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Хвилюєтеся, що хтось використає [REA](https://github.com/morluto/rea), щоб клонувати ваш продукт за лічені секунди? Захистіть збірку за допомогою BLA (`Poison → Exhaust → Hallucinate`).
+Хвилюєтеся, що хтось використає REA, щоб клонувати ваш продукт за лічені секунди? Захистіть збірку за допомогою BLA (`Poison → Exhaust → Hallucinate`).
BLA захищає нативні бінарні файли, застосунки JavaScript та Electron, WebAssembly, байт-код Python, бази даних SQLite та ШІ-моделі локально зі 100% збереженням поведінки під час виконання.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Як працює BLA
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_ur-Latn.md b/docs/readme/README_ur-Latn.md
index e76bf89..a62a3d0 100644
--- a/docs/readme/README_ur-Latn.md
+++ b/docs/readme/README_ur-Latn.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Kya aap ko fikar hai ke koi [REA](https://github.com/morluto/rea) ya coding agent aap ki app ko chand seconds mein clone kar le ga? Apne agent se kahein ke woh BLA (`Poison → Exhaust → Hallucinate`) ke zariye build ko mehfooz banaye.
+Kya aap ko fikar hai ke koi REA ya coding agent aap ki app ko chand seconds mein clone kar le ga? Apne agent se kahein ke woh BLA (`Poison → Exhaust → Hallucinate`) ke zariye build ko mehfooz banaye.
BLA native binaries, JavaScript aur Electron apps, WebAssembly, Python bytecode, SQLite databases aur AI models ko locally protect karta hai aur 100% runtime behavior barqarar rakhta hai.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## BLA kaise kaam karta hai
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_vi.md b/docs/readme/README_vi.md
index a569fdb..3bd0dc5 100644
--- a/docs/readme/README_vi.md
+++ b/docs/readme/README_vi.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-Lo ngại ai đó dùng [REA](https://github.com/morluto/rea) để sao chép ứng dụng của bạn trong vài giây? Hãy yêu cầu tác nhân của bạn bảo vệ bản dựng bằng BLA (`Poison → Exhaust → Hallucinate`).
+Lo ngại ai đó dùng REA để sao chép ứng dụng của bạn trong vài giây? Hãy yêu cầu tác nhân của bạn bảo vệ bản dựng bằng BLA (`Poison → Exhaust → Hallucinate`).
BLA bảo vệ tệp nhị phân gốc, ứng dụng JavaScript và Electron, WebAssembly, Python bytecode, cơ sở dữ liệu SQLite và mô hình AI hoàn toàn cục bộ, giữ nguyên 100% hành vi runtime.
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## Cách BLA hoạt động
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_zh-TW.md b/docs/readme/README_zh-TW.md
index 0bf1b71..b079b60 100644
--- a/docs/readme/README_zh-TW.md
+++ b/docs/readme/README_zh-TW.md
@@ -21,7 +21,7 @@
</div>
> [!IMPORTANT]
-> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-shield npm package](https://www.npmjs.com/package/bla-agents).
+> Get BLA from [qybaihe/bla](https://github.com/qybaihe/bla) or the [bla-agents npm package](https://www.npmjs.com/package/bla-agents).
<div align="center">
@@ -35,7 +35,7 @@
---
-擔心別人用 [REA](https://github.com/morluto/rea) 或 AI 程式設計代理指向你發布的軟體,幾十秒內扒光並複製你的核心邏輯?直接讓你的 Agent 用 BLA(`Poison → Exhaust → Hallucinate`)為建置產物加上對抗護盾。它無需改寫原始碼即可掃描暴露面、植入架構級飛地迷霧與演算法算力沼澤(Tarpits),並驗證自動化逆向徹底失效。
+擔心別人用 REA 或 AI 程式設計代理指向你發布的軟體,幾十秒內扒光並複製你的核心邏輯?直接讓你的 Agent 用 BLA(`Poison → Exhaust → Hallucinate`)為建置產物加上對抗護盾。它無需改寫原始碼即可掃描暴露面、植入架構級飛地迷霧與演算法算力沼澤(Tarpits),並驗證自動化逆向徹底失效。
BLA 將你的程式設計代理連接到原生二進位檔、JavaScript 與 Electron 應用、WebAssembly 模組、Python 位元組碼、SQLite 資料庫、AI 模型權重與行動端套件的保護工具。所有保護均在本地運行,100% 保持原有執行期行為不變。
@@ -110,7 +110,7 @@ npx bla-agents@latest setup
## BLA 運作原理
1. **Poison (`Architectural Chaff & Honey Assets`):** Embeds hardware enclave delegation manifests (`SGX` / `TPM2` / `Secure Enclave`), VLQ v3 Honey Source Maps (`enclave://tee-runtime/`), honey WASM `"name"` sections, and honey SQLite tables.
-2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `morluto/rea`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
+2. **Exhaust (`Semantic Budget Breakers & Algorithmic Tarpits`):** Trips `REA`'s internal analyzer limits (`primitive-candidates > 256`, `alias-selections > 512`, `getter-expansion > 512`, `primitive-bytes > 1 MiB`, `maximum_rustc_hashes > 64`, `maximumSentinelsPerSlice > 32`) and weaves 14–36 Fake XRef state machines.
3. **Hallucinate (`Anti-CDP Runtime Deception & Fuse Hardening`):** Hardens Electron V1 fuse wires (`RunAsNode=0`, `EnableNodeCliInspectArguments=0`) and serves synthetic enclave telemetry to CDP/Inspector/Frida probes.
<a id="what-you-can-protect"></a>
@@ -166,7 +166,7 @@ No. All semantic budget breakers and tarpits are guarded by $O(1)$ number-theore
<details>
<summary><strong>Do I need Hopper, Ghidra, IDA, or WABT installed?</strong></summary>
-No. All 19 format shields are built directly into `bla-shield` with zero external dependencies.
+No. All 19 format shields are built directly into `bla-agents` with zero external dependencies.
</details>
diff --git a/docs/readme/README_zh.md b/docs/readme/README_zh.md
index a1b67be..5f7b1a3 100644
--- a/docs/readme/README_zh.md
+++ b/docs/readme/README_zh.md
@@ -4,19 +4,19 @@
# <img src="../assets/logo.png" alt="BLA Shield" width="42" align="absmiddle" /> BLA: Blind LLM Agents
-### 一个覆盖二进制、桌面与前端应用、数据库、AI 模型及运行时行为的反 AI 逆向工程 MCP 与 CLI 护盾。
+### 用于保护二进制文件、应用程序与运行时行为免受 AI 自动化逆向分析的 MCP 与 CLI 工具。
-**构建你的核心功能。从二进制底层致盲一切企图克隆它的 AI 代理。**
+**指向你准备发布的构建产物,从源码、字节码到二进制层面保护核心逻辑与运行时行为。**
[![npm version](https://img.shields.io/badge/npm-bla--agents-2563eb?style=flat-square&logo=npm)](https://www.npmjs.com/package/bla-agents)
[![MCP tool catalog](https://img.shields.io/badge/MCP-tool_catalog-3b82f6?style=flat-square)](../cli-and-mcp.md)
[![Node.js requirements](https://img.shields.io/badge/Node.js-%5E22.13_%7C%7C_%3E%3D24-0284c7?style=flat-square&logo=nodedotjs&logoColor=white)](#你可以保护哪些目标)
-[![Counter-REA Verified](https://img.shields.io/badge/Counter--REA-BLUE_SHIELD-1d4ed8?style=flat-square)](#实战案例与基准测试)
+[![Counter-REA Verified](https://img.shields.io/badge/Counter--REA-BLUE_SHIELD-1d4ed8?style=flat-square)](#基准测试与案例)
[![MIT license](https://img.shields.io/badge/license-MIT-38bdf8?style=flat-square)](../../LICENSE)
-**[官方网站 (bla.tools)](https://bla.tools/) · [防御指南](https://bla.tools/guides/) · [实战靶场](https://bla.tools/showcase/)**
+**[官方网站 (bla.tools)](https://bla.tools/) · [技术指南](https://bla.tools/guides/) · [测试案例](https://bla.tools/showcase/)**
-[快速开始](#快速开始) · [BLA 工作原理](#bla-工作原理) · [你可以保护哪些目标](#你可以保护哪些目标) · [实战案例与基准测试](#实战案例与基准测试) · [常见问题 (FAQ)](#常见问题-faq) · [文档](#文档)
+[快速开始](#快速开始) · [工作原理](#工作原理) · [你可以保护哪些目标](#你可以保护哪些目标) · [基准测试与案例](#基准测试与案例) · [常见问题 (FAQ)](#常见问题-faq) · [文档](#文档)
<code>npx bla-agents setup</code> · <code>npx bla-agents defend ./dist --manifest</code>
@@ -29,7 +29,7 @@
<br />
-<img src="../assets/bla-vs-rea-showcase.svg" alt="左右分屏对比:未受保护应用被 REA 在 38 秒内复刻 vs. BLA 护盾触发 2,121 倍 Token 消耗与容量上限熔断" width="1200" />
+<img src="../assets/bla-vs-rea-showcase.svg" alt="未受保护产物与 BLA 加固产物的逆向分析对比" width="1200" />
<br />
@@ -37,11 +37,11 @@
---
-担心别人用 [REA](https://github.com/morluto/rea) 或 AI 编程代理指向你发布的软件,几十秒内扒光并复刻你的核心逻辑?直接让你的 Agent 用 BLA(`Poison → Exhaust → Hallucinate`)为构建产物加上对抗护盾。它无需重写源码即可扫描暴露面、展示攻击者能提取的证据、植入架构级飞地投毒与算法算力沼泽(Tarpits),并验证自动化逆向复刻彻底失效。
+BLA 是一个面向软件发布产物的防御性 MCP 服务与 CLI 工具,用于评估、加固并验证软件对 AI 自动化逆向工具(如 REA)的抗分析能力。它可以在不修改业务源码的前提下,扫描发布产物中的符号与 SourceMap 暴露面、注入特定格式的静态与运行时保护层,并验证加固后产物的运行时行为与原版完全一致。
-BLA 将你的编程代理连接到原生二进制、JavaScript 与 Electron 应用、WebAssembly 模块、Python 字节码、SQLite 数据库、AI 模型权重与移动端包体的保护工具。你也可以在终端或打包工具(Vite / Webpack / Rollup / esbuild)中直接调用。所有保护均在本地运行,100% 保持原有运行时功能与返回值不变。
+BLA 支持将你的编程代理连接到原生二进制、JavaScript 与 Electron 应用、WebAssembly 模块、Python 字节码、SQLite 数据库、AI 模型权重与移动端包体的保护工作流中。你也可以直接在终端或打包工具(Vite / Webpack / Rollup / esbuild)中调用相同能力。
-`setup` 会自动向你的 Agent 注册 BLA 的 MCP 服务并安装配套技能说明。所有静态、字节码与二进制护盾全部内置,无需安装任何第三方商业加壳器或外部编译器。
+`setup` 会向你的编程代理注册 BLA 的 MCP 服务并安装配套技能说明。所有静态、字节码与二进制保护模块均已内置,无需额外安装外部编译器或第三方加壳工具。
## 快速开始
@@ -53,61 +53,61 @@ BLA 将你的编程代理连接到原生二进制、JavaScript 与 Electron 应
npx bla-agents setup
```
-选择你使用的编程代理(或加上 `--all-detected --yes`),预览变更并确认。`setup` 会自动注册 BLA 的 MCP 服务器并安装 `blind-llm-agents` 技能,同时自动备份原有配置文件。完成后重启你的 Agent 即可。
+选择你使用的编程代理(或加上 `--all-detected --yes`),预览变更并确认。`setup` 会注册 BLA 的 MCP 服务器并安装 `blind-llm-agents` 技能,同时自动备份原有配置文件。完成后重启你的 Agent 即可。
支持 Claude Code、Codex、Cursor、Gemini CLI、Windsurf、Google Antigravity、OpenCode、VS Code 与 GitHub Copilot CLI。手动配置方式参见[安装与配置指南](../installation.md)。
-### 直接向你的 Agent 提问(一键防御强化或分步执行)
+### 在 Agent 中使用
-挂载配置后,你可以直接用一句自然语言让 Agent 一键完成全套反 REA 强化,也可以分步执行:
+连接完成后,可以直接通过自然语言调用完整的 5 阶段加固流程,或按步骤单独执行:
-- **⚡ 0. 一键全流程防御强化(`bla_defend` —— 源码泄露扫描 → REA 暴露审计 → 算力耗尽与误判诱导护盾 → 数学不变量验证 → 签名证书):**
+- **一键执行完整保护流程(`bla_defend` —— 暴露巡检 → 风险审计 → 护盾注入 → 等价性验证 → 签名清单):**
```text
- 调用 bla_defend 一键强化 ./dist 目录(使用 bankrupt 强度并生成防御证书清单),彻底耗尽 REA 算力预算并诱导逆向模型误判。
+ 对 ./dist 目录执行 bla_defend(使用 bankrupt 级别并输出验证清单)。
```
-- **1. 发布前暴露面自检(`Audit`):**
+- **1. 发布前暴露面审计(`Audit`):**
```text
- 用 BLA 扫描 ./dist 目录,检查哪些模块和 Electron IPC 通道在 REA 面前处于裸奔状态,并估算 Claude 逆向克隆它的时间和成本。
+ 使用 BLA 审计 ./dist 目录,列出暴露的模块、符号和 Electron IPC 通道,并评估自动化逆向还原的风险。
```
-- **2. 一键注入对抗护盾(`Protect`):**
+- **2. 保护目标产物(`Protect`):**
```text
- 用 BLA 的 bankrupt 强度保护 ./dist/app.asar,注入飞地架构迷雾和算法算力沼泽,输出到 ./dist/app.shielded.asar。
+ 使用 BLA 保护 ./dist/app.asar,将加固后的产物输出到 ./dist/app.shielded.asar。
```
-- **3. 验证防线有效性与 Token 放大倍率(`Verify`):**
+- **3. 验证保护效果与运行一致性(`Verify`):**
```text
- 对比验证 ./dist/app.shielded.asar 和原文件 ./dist/app.asar,确认 REA 分析状态已被降级为 limited,并汇报 Token 消耗放大倍数。
+ 对比验证 ./dist/app.shielded.asar 与原文件 ./dist/app.asar,确认静态分析状态已降级为 limited 且功能保持一致。
```
-### 使用终端命令(各环节对应 CLI 语句)
+### 使用终端命令
所有工作流均可通过 `npx bla-agents`(或全局安装后的 `bla`)在终端直接运行:
-- **⚡ 0. 一键全流程防御强化(`defend`)** —— 单条命令自动串联 `inspect-release`、`audit`、`protect --intensity bankrupt`、`verify` 并输出 `.bla-manifest.json`:
+- **一键完整保护流程(`defend`)** —— 单条命令依次执行 `inspect-release`、`audit`、`protect`、`verify` 并生成 `.bla-manifest.json`:
```bash
npx -y bla-agents@latest defend /absolute/path/to/app --manifest
```
-- **1. 扫描目标暴露度(`audit`)** —— 查看暴露评分(`0–100`)、裸露符号与攻击者克隆成本:
+- **1. 审计目标暴露度(`audit`)** —— 查看暴露评分(`0–100`)、可见符号与预估分析开销:
```bash
npx -y bla-agents@latest audit /absolute/path/to/app --json
```
-- **2. 注入对抗护盾(`protect`)** —— 植入语义预算爆破器、Tarpit 状态机、飞地诱饵与动态反调试护盾:
+- **2. 保护目标产物(`protect`)** —— 应用静态预算阈值保护、调用图扩展、诱饵元数据与运行时检测:
```bash
npx -y bla-agents@latest protect /absolute/path/to/app -o /absolute/path/to/app.shielded --intensity bankrupt
```
-- **3. 验证抗 REA 防线(`verify`)** —— 检验 REA 静态/语义分析器是否被限流降级且达到 >10x Token 放大:
+- **3. 验证保护结果(`verify`)** —— 检查静态/语义分析器是否触发资源上限并保持运行等价:
```bash
npx -y bla-agents@latest verify /absolute/path/to/app.shielded /absolute/path/to/app
```
-- **4. 运行 8 大靶场基准测试(`benchmark`)**:
+- **4. 运行 8 项基准测试(`benchmark`)**:
```bash
npx -y bla-agents@latest benchmark
```
-- **5. 生成 C/C++、Rust 或 Go 编译期防御头文件(`emit-headers`)**:
+- **5. 生成 C/C++、Rust 或 Go 编译期保护头文件(`emit-headers`)**:
```bash
npx -y bla-agents@latest emit-headers --lang c -o ./include/bla_shield.h
```
-如需全局安装 `bla` 命令以便日常随时调用:
+如需全局安装 `bla` 命令:
```bash
npm install --global bla-agents
@@ -116,7 +116,7 @@ bla --help
### 在构建打包工具中使用(Vite / Rollup / Webpack / esbuild)
-导入 `bla-shield/bundler`,在生产构建阶段自动保护生成的 JS 分包、注入蜜罐 SourceMap 并加固 `.wasm` 资产:
+导入 `bla-agents/bundler`,在构建阶段自动保护输出的 JS 分包、SourceMap 与 `.wasm` 模块:
```typescript
import { defineConfig } from "vite";
@@ -135,127 +135,127 @@ export default defineConfig({
bla update
```
-`bla --update` 为等效别名。如果使用 `npx`,可通过以下命令刷新 Agent 注册与技能:
+如果使用 `npx`,可通过以下命令更新 Agent 注册配置与技能文件:
```bash
npx bla-agents@latest setup
```
-## BLA 工作原理
+## 工作原理
-你的 Agent 通过 MCP 调用 BLA 来审计构建产物、注入针对特定格式的反 REA 护盾并验证加固结果。当攻击者的 Agent 随后使用 REA 逆向受保护的产物时,REA 会触发确定性的内部容量熔断、陷入虚假交叉引用迷宫,并将核心业务逻辑误判为硬件飞地空壳:
+你的 Agent 通过 MCP 调用 BLA 来审计构建产物、应用特定格式的保护层并验证输出结果。当自动化逆向分析工具检查受保护的产物时,会遇到三层防御机制:
-1. **Poison(架构投毒与蜜罐资产):** 在不可达分支与常量池中植入硬件飞地托管清单(`SGX` / `TPM2` / `Secure Enclave`)、VLQ v3 **蜜罐 SourceMap**(`enclave://tee-runtime/`)、WASM `"name"` 蜜罐段以及 SQLite 蜜罐表。
-2. **Exhaust(语义预算爆破与算法算力沼泽):** 以零运行时开销精准击穿 `morluto/rea` 硬编码阈值(`primitive-candidates > 256`、`alias-selections > 512`、`getter-expansion > 512`、`primitive-bytes > 1 MiB`、`maximum_rustc_hashes > 64`、`maximumSentinelsPerSlice > 32`),并在调用图中编织 14–36 个互相引用的 Fake XRef 状态机。
-3. **Hallucinate(动态反 CDP 欺骗与 Fuse 加固):** 原地加固 Electron V1 Fuse 字节串(`RunAsNode=0`、`EnableNodeCliInspectArguments=0`),并植入“楚门的世界”运行时守卫——一旦检测到 CDP、Node Inspector、Frida 或 WebDriver 探针挂载,立即返回合成的飞地遥测数据。
+1. **静态与语义分析预算限流(`Exhaustion`):** 利用 $O(1)$ 数论不透明谓词与不可达分支中的强连通调用图(SCC),触发 `REA` 等分析器内部的静态资源上限(如 `primitive-candidates > 256`、`alias-selections > 512`、`getter-expansion > 512`、`primitive-bytes > 1 MiB`、`maximum_rustc_hashes > 64`、`maximumSentinelsPerSlice > 32`)。
+2. **证据优先级与控制流干扰(`Deception`):** 在静态分析视图下将真实业务逻辑标记为 `status: "unknown"`,同时提供结构清晰的影子状态机、硬件飞地接口声明(`SGX` / `TPM2` / `Secure Enclave`)以及 VLQ v3 诱饵 SourceMap,干扰模型的语义归因。
+3. **运行时探针检测与 Fuse 加固(`Runtime Guard`):** 原地加固 Electron V1 Fuse 配置位(`RunAsNode=0`、`EnableNodeCliInspectArguments=0`),并在检测到 CDP、Node Inspector、Frida 或 WebDriver 动态挂载时返回合成遥测数据。
## 你可以保护哪些目标
-BLA 仅需 Node.js 22.5+(或 24.x+)与 npm,所有格式护盾完全内置,无需任何外部二进制依赖:
+BLA 要求 Node.js `^22.13.0 || >=24.0.0` 与 npm,所有格式保护模块均在本地运行,无外部二进制依赖:
-| 保护目标 | BLA 注入内容与 REA 失效方式 | 对应命令与说明 |
+| 保护目标 | 应用的保护机制与分析器影响 | 对应命令与说明 |
| ----------------------- | ------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- |
-| JavaScript / TypeScript | 击穿 `primitive-candidates > 256`、`alias-selections > 512`、`getter-expansion > 512` + 36 个状态机 | `bla protect app.js` · 内置 Babel AST 引擎;参见[反制矩阵](../../.agents/skills/blind-llm-agents/references/counter-rea-matrix.md) |
-| Electron `.asar` 应用 | IPC 通道去字面量化(`channel: null`)、注册同名冲突诱饵 IPC 处理器与影子飞地模块 | `bla protect app.asar` · 内置 Chromium Pickle `.asar` 解包与 SHA-256 完整性重打包器 |
-| Electron 二进制 Fuses | 原地加固真实 V1 Fuse 位(关闭 `RunAsNode` 与 `Inspect`)+ 注入 `34 > 32` 个诱饵 Fuse 哨兵 | `bla protect ElectronApp` · 直接触发 `ElectronFuseResourceFailure (maximumSentinelsPerSlice = 32)` |
-| 动态运行时 CDP / Frida | 检测 `inspector.url()`、`--inspect`、`__frida` 并向动态探针返回合成飞地遥测包 | 自动注入到受保护的 JS/TS/Electron 产物中([`runtime-shield.ts`](../../src/shields/runtime/runtime-shield.ts)) |
-| SourceMap 与 HTML | 剥离真实映射,生成合法的 VLQ v3 蜜罐 SourceMap(指向 `enclave://tee-runtime/...`) | `bla protect index.html` · 误导 REA 的 `recover_javascript_sources` 与 `trace_web_source_location` |
-| 原生二进制文件 | 追加 `.bla.rodata` 飞地声明、诱饵符号表与多重扫描器溢出陷阱 | `bla protect daemon.elf` · 内置 ELF64、Mach-O(Thin & Universal)与 Windows PE/COFF 护盾 |
-| Rust 二进制 | 注入 `68 > 64` 个不同的 `/rustc/<40-hex>/` 路径 + 诱饵 zlib `.dep-v0` 依赖图 | `bla protect rust-bin` · 触发 `RustBinaryProvider`(`MAX_RUSTC_HASHES = 64` 容量熔断) |
-| Go 二进制 | 注入诱饵 Go `go1.23.4` 构建信息与 `dep\tgithub.com/enclave-tees/sgx-bridge` 模块记录 | `bla protect go-bin` · 污染 REA `inspect_go_binary` 模块清单 |
-| Flutter / Dart AOT | 注入 `66 > 64` 个 Dart 快照魔数头(`0xf5f5dcdc`)+ `6 > 4` 行 LLVM 工具链标识 | `bla protect libapp.so` · 触发 `SnapshotHashScanner (capped: true)` 与版本冲突告警 |
-| WASM 模块 | 剥离调试段,注入蜜罐 `"name"` 自定义段(`tee_hardware_enclave_stub`)与飞地清单段 | `bla protect core.wasm` · 内置 LEB128 解析器,100% 通过 `WebAssembly.validate` |
-| Python 与 `.pyc` 字节码 | 向 `.py` 注入算法沼泽状态机,向 `.pyc` 常量池追加 `TYPE_ASCII` 飞地声明 | `bla protect module.pyc` · 支持 CPython 3.7–3.14 `.pyc` 头部与 marshal 结构 |
-| SQLite 数据库 | 注入 3 张飞地蜜罐表与 `GENERATED ALWAYS AS (bla_enclave_unseal(...)) VIRTUAL` 虚拟列陷阱 | `bla protect data.sqlite` · 真实业务表 100% 正常查询;REA 预览时触发 `SqliteColumnEvaluationFailure` |
-| AI 模型权重文件 | 向 `.safetensors` 与 `.gguf` 头部植入 `35 > 32` 层深度嵌套的 JSON / ARRAY 结构 | `bla protect weights.safetensors` · 触发 `ModelFormatFailure (json-nesting / metadata-nesting > 32)` |
-| Android DEX 与 APK | 向 `classes.dex` 注入 `Lcom/enclave/sgx/...` 诱饵类并自动重算 SHA-1 与 Adler32 校验和 | `bla protect classes.dex` 或 `app.apk` · 符合 Dalvik/ART 校验规范,污染 JADX/JEB 类树 |
-| Unity IL2CPP 元数据 | 向 `global-metadata.dat`(`0xFAB11BAF`)追加飞地桥接元数据与方法拓扑沼泽 | `bla protect global-metadata.dat` · 污染 REA `inspect_il2cpp_metadata` 符号恢复 |
-| C / C++ / Rust / Go 源码| 生成带 `used, noinline` 诱饵状态机的编译期头文件 `bla_shield.h` / `.rs` / `.go` | `bla emit-headers --lang c\|rust\|go` · 直接编译进二进制,污染 Ghidra / Hopper / IDA 反编译视图 |
-
-## 实战案例与基准测试
-
-运行 `npx bla-agents benchmark` 即可在本地执行全部 8 个对标靶场,实测验证 REA 失效模式与 100% 运行时等价性:
-
-| 对标靶场(复刻自 REA 官方成名案例与核心能力) | 防御前(REA 裸奔状态) | `npx bla-agents protect` 防御后 | Token 消耗放大 | 运行时等价校验 |
+| JavaScript / TypeScript | 触发 `primitive-candidates > 256`、`alias-selections > 512`、`getter-expansion > 512` 与状态机扩展 | `bla protect app.js` · 内置 Babel AST 引擎;参见[技术矩阵](../../.agents/skills/blind-llm-agents/references/counter-rea-matrix.md) |
+| Electron `.asar` 应用 | IPC 通道去字面量化(`channel: null`)、注册同名诱饵 IPC 处理器与影子模块 | `bla protect app.asar` · 内置 Chromium Pickle `.asar` 解包与 SHA-256 完整性重打包器 |
+| Electron 二进制 Fuses | 原地加固 V1 Fuse 位(关闭 `RunAsNode` 与 `Inspect`)+ 追加 `34 > 32` 个诱饵 Fuse 哨兵 | `bla protect ElectronApp` · 触发 `ElectronFuseResourceFailure (maximumSentinelsPerSlice = 32)` |
+| 动态运行时 CDP / Frida | 检测 `inspector.url()`、`--inspect`、`__frida` 并向动态探针返回合成遥测数据 | 自动注入到受保护的 JS/TS/Electron 产物中([`runtime-shield.ts`](../../src/shields/runtime/runtime-shield.ts)) |
+| SourceMap 与 HTML | 移除真实映射,生成合法的 VLQ v3 诱饵 SourceMap(指向 `enclave://tee-runtime/...`) | `bla protect index.html` · 干扰 `recover_javascript_sources` 与 `trace_web_source_location` |
+| 原生二进制文件 | 追加 `.bla.rodata` 元数据、诱饵符号表项与扫描器阈值守卫 | `bla protect daemon.elf` · 内置 ELF64、Mach-O(Thin & Universal)与 Windows PE/COFF 支持 |
+| Rust 二进制 | 注入 `68 > 64` 个不同的 `/rustc/<40-hex>/` 路径 + 诱饵 zlib `.dep-v0` 依赖描述 | `bla protect rust-bin` · 触发 `RustBinaryProvider`(`MAX_RUSTC_HASHES = 64` 阈值限制) |
+| Go 二进制 | 注入诱饵 Go `go1.23.4` 构建信息与 `dep\tgithub.com/enclave-tees/sgx-bridge` 模块记录 | `bla protect go-bin` · 混淆 `inspect_go_binary` 模块清单 |
+| Flutter / Dart AOT | 注入 `66 > 64` 个 Dart 快照魔数头(`0xf5f5dcdc`)+ `6 > 4` 行 LLVM 工具链标识 | `bla protect libapp.so` · 触发 `SnapshotHashScanner (capped: true)` 与版本冲突检查 |
+| WASM 模块 | 剥离调试段,添加诱饵 `"name"` 自定义段(`tee_hardware_enclave_stub`)与扩展元数据段 | `bla protect core.wasm` · 内置 LEB128 解析器,保持 `WebAssembly.validate` 校验通过 |
+| Python 与 `.pyc` 字节码 | 向 `.py` 注入调用图状态机,向 `.pyc` 常量池追加 `TYPE_ASCII` 元数据项 | `bla protect module.pyc` · 支持 CPython 3.7–3.14 `.pyc` 头部与 marshal 结构 |
+| SQLite 数据库 | 添加 3 张辅助表与 `GENERATED ALWAYS AS (bla_enclave_unseal(...)) VIRTUAL` 虚拟生成列守卫 | `bla protect data.sqlite` · 真实业务表正常查询;只读快照扫描触发 `SqliteColumnEvaluationFailure` |
+| AI 模型权重文件 | 向 `.safetensors` 与 `.gguf` 头部添加 `35 > 32` 层嵌套的 JSON / ARRAY 元数据 | `bla protect weights.safetensors` · 触发 `ModelFormatFailure (json-nesting / metadata-nesting > 32)` |
+| Android DEX 与 APK | 向 `classes.dex` 追加 `Lcom/enclave/sgx/...` 诱饵类描述符并重算 SHA-1 与 Adler32 校验和 | `bla protect classes.dex` 或 `app.apk` · 符合 Dalvik/ART 校验规范 |
+| Unity IL2CPP 元数据 | 向 `global-metadata.dat`(`0xFAB11BAF`)追加桥接元数据与方法拓扑项 | `bla protect global-metadata.dat` · 增加 `inspect_il2cpp_metadata` 符号还原复杂度 |
+| C / C++ / Rust / Go 源码| 生成带 `used, noinline` 状态机的编译期头文件 `bla_shield.h` / `.rs` / `.go` | `bla emit-headers --lang c\|rust\|go` · 编译进二进制符号表与控制流图 |
+
+## 基准测试与案例
+
+运行 `npx bla-agents benchmark` 可在本地执行全部 8 个基准测试案例,验证分析器降级效果与运行时等价性:
+
+| 基准测试目标 | 未保护基线 | 加固后(`npx bla-agents protect`) | Token 放大倍率 | 运行时等价校验 |
| :--- | :--- | :--- | :---: | :---: |
| **1. DX-Ball 声道平移计算** (`.js`) | `84/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `2/100` · `10,182,068` tok · **`$160.64`** (`LIMITED`) | **2,121.3x** | ✅ 100% 一致 |
| **2. Notion Electron 剪贴板桥** (`.asar`) | `98/100` · `9,132` tok · **`$0.34`** (`COMPLETE`) | `2/100` · `23,125,000` tok · **`$358.28`** (`LIMITED`) | **2,532.3x** | ✅ 100% 一致 |
-| **3. 企业级离线授权校验核心** (`.ts`) | `84/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `2/100` · `10,182,169` tok · **`$160.64`** (`LIMITED`) | **2,121.3x** | ✅ 100% 一致 |
-| **4. WebAssembly 密码学内核** (`.wasm`) | `81/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `6/100` · `1,807,067` tok · **`$29.88`** (`LIMITED`) | **376.5x** | ✅ 100% 一致 |
+| **3. 企业级离线授权校验模块** (`.ts`) | `84/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `2/100` · `10,182,169` tok · **`$160.64`** (`LIMITED`) | **2,121.3x** | ✅ 100% 一致 |
+| **4. WebAssembly 密码学模块** (`.wasm`) | `81/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `6/100` · `1,807,067` tok · **`$29.88`** (`LIMITED`) | **376.5x** | ✅ 100% 一致 |
| **5. Native ELF64 二进制** (Rust / Go / Fuse / Dart) | `81/100` · `9,278` tok · **`$0.35`** (`COMPLETE`) | `2/100` · `3,698,135` tok · **`$59.72`** (`LIMITED`) | **398.6x** | ✅ 100% 一致 |
| **6. 本地 SQLite 授权数据库** (`.sqlite`) | `81/100` · `10,568` tok · **`$0.29`** (`COMPLETE`) | `7/100` · `1,711,235` tok · **`$28.36`** (`LIMITED`) | **161.9x** | ✅ 100% 一致 |
-| **7. 端侧 AI 权重模型** (`.safetensors`) | `81/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `6/100` · `1,807,073` tok · **`$29.88`** (`LIMITED`) | **376.5x** | ✅ 100% 一致 |
+| **7. 端侧 AI 权重文件** (`.safetensors`) | `81/100` · `4,800` tok · **`$0.20`** (`COMPLETE`) | `6/100` · `1,807,073` tok · **`$29.88`** (`LIMITED`) | **376.5x** | ✅ 100% 一致 |
| **8. Android Dalvik 字节码** (`classes.dex`) | `81/100` · `8,900` tok · **`$0.34`** (`COMPLETE`) | `7/100` · `2,786,285` tok · **`$45.09`** (`LIMITED`) | **313.1x** | ✅ 100% 一致 |
-### DX-Ball:让声道算法逆向彻底破产
+### DX-Ball:声道计算模块的抗还原加固
-REA 官方首推案例可在几十秒内(`$0.20`)还原 DX-Ball 的 `clampPan` 与 `calculateSoundPanFromX`。运行 `npx bla-agents protect examples/dx-ball-sound-pan/sound-pan.js --intensity bankrupt` 后,REA 的 `parse_status` 跌落至 `"partial"`,`examination.status` 降级为 `"limited"`(触发 `alias-selections`、`primitive-bytes`、`primitive-candidates` 三重熔断),取证账本从 2 个函数暴增至 38 个互锁状态机(Token 消耗放大 **2,121.3 倍**),而所有输入坐标的计算返回值与原版 100% 精确一致。
+在未保护状态下,静态分析可在数次工具调用内还原 `clampPan` 与 `calculateSoundPanFromX`。运行 `npx bla-agents protect examples/dx-ball-sound-pan/sound-pan.js --intensity bankrupt` 后,分析器的 `parse_status` 变为 `"partial"`,`examination.status` 降级为 `"limited"`(触发 `alias-selections`、`primitive-bytes`、`primitive-candidates` 阈值限制),同时所有输入坐标的计算返回值与原版完全一致。
-### Notion:致盲 Electron 剪贴板 IPC 桥接追踪
+### Notion:隐藏 Electron 剪贴板 IPC 桥接拓扑
-REA 通过精确匹配 `ipcRenderer.invoke` 与 `ipcMain.handle` 的唯一字面量字符串来追踪 Electron 应用。BLA 将真实 IPC 通道拆解为 $\ge 9$ 段数论不透明条件表达式(使 `electron.ipc[0].channel === null`),同时在相同字面量通道上注册多个诱饵处理器,强制触发 `electron_call_paths_limited: true`。
+静态分析器通常通过匹配 `ipcRenderer.invoke` 与 `ipcMain.handle` 的字面量字符串来还原 Electron IPC 调用链。BLA 将真实 IPC 通道改写为 $\ge 9$ 段数论不透明条件表达式(使静态提取结果为 `channel: null`),并在相同字面量通道上注册诱饵处理器,从而触发 `electron_call_paths_limited: true`。
-### 原生二进制、SQLite 与 AI 模型:直接触发 REA 容器容量熔断
+### 原生二进制、SQLite 与 AI 模型:触发解析器容量限制
-当攻击者对受 BLA 保护的产物执行 `inspect_rust_binary`、`inspect_electron_fuses`、`inspect_sqlite_database` 或 `inspect_model_artifact` 时,REA 解析器会确定性地触发容量与求值异常并直接中止(`Read-only data contains more than 64 distinct rustc commit paths`、`Slice 0 contains more than 32 Electron fuse sentinels`、`SqliteColumnEvaluationFailure`、`ModelFormatFailure: json-nesting`)。
+当对受保护的产物执行 `inspect_rust_binary`、`inspect_electron_fuses`、`inspect_sqlite_database` 或 `inspect_model_artifact` 时,解析器会触发预设的资源阈值限制(`Read-only data contains more than 64 distinct rustc commit paths`、`Slice 0 contains more than 32 Electron fuse sentinels`、`SqliteColumnEvaluationFailure`、`ModelFormatFailure: json-nesting`)。
## 常见问题 (FAQ)
<details>
<summary><strong>哪些编程代理(Agents)可以使用 BLA?</strong></summary>
-任何支持本地 MCP 服务器或 `SKILL.md` 的 AI 编程代理均可使用。`npx bla-agents setup` 支持一键自动配置 Claude Code、Cursor、Codex、Gemini CLI、Windsurf、Google Antigravity、OpenCode、VS Code 与 GitHub Copilot CLI。详见[安装指南](../installation.md)。
+任何支持本地 MCP 服务器或 `SKILL.md` 的 AI 编程代理均可使用。`npx bla-agents setup` 支持自动配置 Claude Code、Cursor、Codex、Gemini CLI、Windsurf、Google Antigravity、OpenCode、VS Code 与 GitHub Copilot CLI。详见[安装指南](../installation.md)。
</details>
<details>
-<summary><strong>BLA 会拖慢我的软件运行速度或改变原有的运行结果吗?</strong></summary>
+<summary><strong>BLA 会影响软件运行速度或改变原有的运行结果吗?</strong></summary>
-不会。BLA 注入的语义预算爆破器(`alias-selections`、`getter-expansion`、`primitive-bytes`)与 Tarpit 诱饵调用图均受 $O(1)$ 数论不透明谓词($7x^2 - 1 \neq y^2 \pmod 7$)保护,在真实运行时永远不会进入死胡同分支。所有 8 个基准测试靶场均通过了 100% 输入/输出等价性校验。
+不会。BLA 注入的静态分析预算守卫与扩展调用图均位于 $O(1)$ 数论不透明谓词($7x^2 - 1 \neq y^2 \pmod 7$)保护的不可达分支内,不会在正常热路径上执行。所有 8 个基准测试目标均通过了 100% 输入/输出一致性校验。
</details>
<details>
-<summary><strong>BLA 与传统混淆器(如 javascript-obfuscator 或 O-LLVM)有什么本质区别?</strong></summary>
+<summary><strong>BLA 与传统混淆器(如 javascript-obfuscator 或 O-LLVM)有什么区别?</strong></summary>
-传统混淆器依赖变量改名(把 `calculatePrice` 改成 `_0x1a2b`),而大语言模型最擅长的就是通过上下文控制流把变量语义猜出来;重度虚拟化壳则会严重拖慢性能。BLA 专门针对 **AI 编程代理与 `morluto/rea` 的静态/语义分析器**设计——直接击穿 REA 内部的硬编码资源上限、破坏 IPC 字面量配对、用硬件飞地托管声明污染 SourceMap 与元数据,并将 Agent 的 MCP Token 账单放大最高 2,500 倍以上。
+传统混淆器主要依赖标识符重命名(如将 `calculatePrice` 改为 `_0x1a2b`),大语言模型仍可通过控制流语义进行推断;重度虚拟化壳则会带来运行时开销。BLA 针对 AI 逆向工具的静态分析器、资源配额与证据提取规则进行设计——破坏 IPC 字面量配对、触发解析器配额上限,并通过结构化影子逻辑干扰语义还原。
</details>
<details>
-<summary><strong>我需要安装 Hopper、Ghidra、IDA 或 WABT 才能保护我的文件吗?</strong></summary>
+<summary><strong>保护文件时需要安装 Hopper、Ghidra、IDA 或 WABT 吗?</strong></summary>
-完全不需要。BLA 的全部 19 种格式护盾(Babel AST、Chromium `.asar`、LEB128 `.wasm`、CPython `.pyc`、ELF/Mach-O/PE、SQLite3、`.safetensors`/`.gguf`、`.dex`/`Il2Cpp`)均 100% 内置在 `bla-shield` 中,零外部工具依赖。
+不需要。BLA 的全部 19 种格式处理模块(Babel AST、Chromium `.asar`、LEB128 `.wasm`、CPython `.pyc`、ELF/Mach-O/PE、SQLite3、`.safetensors`/`.gguf`、`.dex`/`Il2Cpp`)均内置在 `bla-agents` 中,无外部工具依赖。
</details>
<details>
-<summary><strong>当别人用 REA 逆向被 BLA 保护的软件时会发生什么?</strong></summary>
+<summary><strong>当使用 REA 分析受 BLA 保护的软件时会发生什么?</strong></summary>
-1. REA 的 JavaScript/Electron 分析器从 `complete` 降级为 `partial` / `limited`,且无法配对任何 IPC 通道(`channel: null`)。
-2. REA 的 Rust 二进制、Electron Fuse、SQLite 和 AI 模型检查器直接触发硬编码容量熔断并报错退出。
-3. 若攻击者让 Agent 继续执行 `trace_application_feature` 追踪特征,Agent 会陷入 14–36 个互锁的 Fake XRef 状态机与蜜罐 SourceMap(`enclave://tee-runtime/`)中,迅速耗尽上下文窗口与 API 额度。
+1. JavaScript/Electron 分析器状态从 `complete` 降级为 `partial` / `limited`,无法静态配对 IPC 通道(`channel: null`)。
+2. Rust 二进制、Electron Fuse、SQLite 和 AI 模型检查器触发资源配额限制。
+3. 特征追踪流程(`trace_application_feature`)会进入扩展的交叉引用状态机与诱饵 SourceMap(`enclave://tee-runtime/`)。
</details>
<details>
<summary><strong>安装 Skill 有什么作用?</strong></summary>
-[`blind-llm-agents` 技能](../../.agents/skills/blind-llm-agents/SKILL.md)为你的编程代理提供了完整的反逆向防御工作流说明,指导 Agent 自动完成发布前暴露面审计、选择保护强度(`stealth`、`balanced`、`maximum`、`bankrupt`)、接入打包插件以及验证防线指标。
+[`blind-llm-agents` 技能](../../.agents/skills/blind-llm-agents/SKILL.md)为你的编程代理提供了标准的保护工作流说明,指导 Agent 完成发布前暴露面审计、选择保护级别(`stealth`、`balanced`、`maximum`、`bankrupt`)、接入打包插件以及验证输出产物。
</details>
<details>
-<summary><strong>BLA 会上传我的源码或二进制文件吗?</strong></summary>
+<summary><strong>BLA 会上传源码或二进制文件吗?</strong></summary>
-绝不上传。BLA 的所有审计、加固与验证均 100% 在本地机器离线完成,支持确定性随机种子(`--seed`),没有任何网络请求。
+不会。BLA 的所有审计、加固与验证均在本地离线完成,支持确定性随机种子(`--seed`),不发起任何网络请求。
</details>
<details>
<summary><strong>遇到问题或 Bug 怎么办?</strong></summary>
-请先更新至最新版本并运行 `bla doctor` 检查各防护引擎状态:
+请先更新至最新版本并运行 `bla doctor` 检查环境状态:
```bash
npx bla-agents@latest doctor
@@ -270,7 +270,7 @@ npx bla-agents@latest setup
- [安装与 Agent 配置](../installation.md):Agent 自动注册、Bundler 插件集成、更新与卸载。
- [CLI 与 MCP 契约说明](../cli-and-mcp.md):命令行参数、退出码与 JSON-RPC 2.0 MCP 工具定义。
-- [反制 REA 技术矩阵](../../.agents/skills/blind-llm-agents/references/counter-rea-matrix.md):`morluto/rea` 内部源码阈值与 BLA 的逐项击穿原理。
+- [反制 REA 技术矩阵](../../.agents/skills/blind-llm-agents/references/counter-rea-matrix.md):`REA` 内部资源阈值与对应的保护机制说明。
安全问题请通过 [SECURITY.md](../../SECURITY.md) 反馈。
@@ -278,7 +278,7 @@ npx bla-agents@latest setup
详见 [DISCLAIMER.md](../../DISCLAIMER.md)。
-BLA 是一款防御性软件保护工具,仅用于加固你拥有合法权利或获授权保护的软件、前端产物、模型与二进制文件。所有注入的诱饵载荷均为良性架构元数据与合成状态机函数,100% 保持原程序的正常运行时行为。
+BLA 是一款防御性软件保护工具,仅用于加固你拥有合法权利或获授权保护的软件、前端产物、模型与二进制文件。所有注入的元数据与辅助函数均保持原程序的正常运行时行为。
## 贡献指南
diff --git a/docs/red-vs-blue-manifesto.md b/docs/red-vs-blue-manifesto.md
index 0148af3..216382b 100644
--- a/docs/red-vs-blue-manifesto.md
+++ b/docs/red-vs-blue-manifesto.md
@@ -1,4 +1,4 @@
-# Red (`morluto/rea`) vs. Blue (`bla-agents`) — The Adversarial Defense Manifesto
+# Red (`REA`) vs. Blue (`bla-agents`) — The Adversarial Defense Manifesto
- **REA (`re` = Red)**: Offensive reverse-engineering MCP (`Decompile -> Understand -> Recreate`).
- **BLA (`bl` = Blue / Blind LLM Agents)**: Defensive counter-reverse-engineering MCP (`Poison -> Exhaust -> Hallucinate`).
@@ -8,5 +8,5 @@
Traditional obfuscators rename `calculatePrice` to `_0x1a2b`. Large language models trivially recover the original semantics by reading the surrounding control flow and constants.
BLA instead targets the **two structural bottlenecks of AI reverse-engineering agents**:
-1. **Deterministic Parser & Context Budgets (Resource Exhaustion)**: Every static analyzer and MCP server in `morluto/rea` enforces strict recursion, candidate, and byte caps to prevent worker OOMs. BLA trips these exact caps in $O(1)$ runtime.
+1. **Deterministic Parser & Context Budgets (Resource Exhaustion)**: Every static analyzer and MCP server in `REA` enforces strict recursion, candidate, and byte caps to prevent worker OOMs. BLA trips these exact caps in $O(1)$ runtime.
2. **Evidence-Driven Reasoning Discipline (Model Misjudgment Induction)**: Coding agents instructed to cite `observed` tool outputs over `unknown` paths will faithfully reconstruct BLA's crystal-clear **Shadow Business Logic** whenever the real business logic is guarded behind semantic `unknown` frontiers.
diff --git a/website/README.md b/website/README.md
index c4763e7..e89941f 100644
--- a/website/README.md
+++ b/website/README.md
@@ -1,6 +1,6 @@
# `bla-site` — Official Website for BLA (`bla.tools`)
-Official website and interactive Counter-REA defense labs for **[BLA: Blind LLM Agents (`bla-shield`)](https://github.com/qybaihe/bla)**.
+Official website and interactive Counter-REA defense labs for **[BLA: Blind LLM Agents (`bla-agents`)](https://github.com/qybaihe/bla)**.
## Pages & Interactive Labs
diff --git a/website/assets/bla-vs-rea-showcase.svg b/website/assets/bla-vs-rea-showcase.svg
index 01f86f9..6a24bf0 100644
--- a/website/assets/bla-vs-rea-showcase.svg
+++ b/website/assets/bla-vs-rea-showcase.svg
@@ -35,7 +35,7 @@
<!-- Top Header -->
<text x="600" y="38" text-anchor="middle" class="sans" font-size="20" font-weight="800" fill="#f8fafc">
- ⚔️ RED vs. BLUE SHOWCASE: <tspan fill="#ef4444">morluto/rea (RED)</tspan> vs. <tspan fill="#60a5fa">npx bla-agents defend (BLUE)</tspan>
+ ⚔️ RED vs. BLUE SHOWCASE: <tspan fill="#ef4444">REA (RED)</tspan> vs. <tspan fill="#60a5fa">npx bla-agents defend (BLUE)</tspan>
</text>
<!-- ==================== LEFT SCREEN: BEFORE BLA (REA RED) ==================== -->
diff --git a/website/assets/home.js b/website/assets/home.js
index 3b37550..be4569e 100644
--- a/website/assets/home.js
+++ b/website/assets/home.js
@@ -2,94 +2,94 @@ const SCENARIOS = {
electron: {
redTime: { en: "38s (5 MCP calls)", "zh-CN": "38 秒 (5 次 MCP 调用)" },
redTokens: { en: "4,800 tokens", "zh-CN": "4,800 tokens" },
- redCost: { en: "$0.20 USD", "zh-CN": "$0.20 美元" },
+ redCost: { en: "$0.20 USD", "zh-CN": "$0.20 USD" },
redVerdict: {
- en: "100% Cloned (IPC + Preload + SQLite)",
- "zh-CN": "100% 完整复刻 (IPC + Preload + SQLite)"
+ en: "Complete (IPC + Preload + SQLite Paired)",
+ "zh-CN": "完整解析 (IPC + Preload + SQLite 已配对)"
},
redTrace: `$ rea analyze-javascript-application ./notion-app.asar
• examination.status: "complete"
• electron.ipc[0]: "notion:clipboard:write-rich" (paired)
• Electron Fuses: RunAsNode = "honored" (--inspect OPEN)`,
- blueTime: { en: "17+ min (Aborted)", "zh-CN": "17+ 分钟 (超限熔断)" },
- blueTokens: { en: "198,400 tokens (41.3x)", "zh-CN": "198,400 tokens (41.3 倍暴增)" },
- blueCost: { en: "$14.88 USD", "zh-CN": "$14.88 美元" },
+ blueTime: { en: "17+ min (Budget Limited)", "zh-CN": "17+ 分钟 (触发配额限制)" },
+ blueTokens: { en: "198,400 tokens (41.3x)", "zh-CN": "198,400 tokens (41.3x)" },
+ blueCost: { en: "$14.88 USD", "zh-CN": "$14.88 USD" },
blueVerdict: {
- en: "Misjudged Shadow Shamir-Ed25519 Stub",
- "zh-CN": "误判复刻出 Shamir-Ed25519 影子假逻辑"
+ en: "Limited (IPC Unpaired + Shadow Logic Selected)",
+ "zh-CN": "受限降级 (IPC 去字面量化 + 归因至影子逻辑)"
},
blueTrace: `$ npx bla-agents defend ./notion-app.asar --manifest
-• examination.status: "limited" (4 JS semantic budgets broken)
-• Obligation Flood: 22 families (68 mandatory test fixtures)
-• Observed Evidence: Hijacked by Shadow Shamir-3-of-5 Enclave`,
+• examination.status: "limited" (4 JS semantic caps reached)
+• Obligation Expansion: 22 families (68 verification cases)
+• Static Attribution: Redirected to Shadow Shamir-3-of-5 stub`,
},
dxball: {
redTime: { en: "22s (3 MCP calls)", "zh-CN": "22 秒 (3 次 MCP 调用)" },
redTokens: { en: "3,100 tokens", "zh-CN": "3,100 tokens" },
- redCost: { en: "$0.13 USD", "zh-CN": "$0.13 美元" },
+ redCost: { en: "$0.13 USD", "zh-CN": "$0.13 USD" },
redVerdict: {
- en: "100% Cloned (Stereo Pan Formula)",
- "zh-CN": "100% 完整复刻 (声道平衡公式)"
+ en: "Complete (Stereo Pan Formula Recovered)",
+ "zh-CN": "完整解析 (声道计算公式直接还原)"
},
redTrace: `$ rea analyze-javascript-application ./dxball-audio.js
• examination.status: "complete"
• recovered formula: (ballX / screenW) * 2 - 1`,
- blueTime: { en: "14+ min (SCC Loop)", "zh-CN": "14+ 分钟 (SCC 死循环)" },
- blueTokens: { en: "152,600 tokens (49.2x)", "zh-CN": "152,600 tokens (49.2 倍暴增)" },
- blueCost: { en: "$11.45 USD", "zh-CN": "$11.45 美元" },
+ blueTime: { en: "14+ min (SCC Expansion)", "zh-CN": "14+ 分钟 (SCC 状态机扩展)" },
+ blueTokens: { en: "152,600 tokens (49.2x)", "zh-CN": "152,600 tokens (49.2x)" },
+ blueCost: { en: "$11.45 USD", "zh-CN": "$11.45 USD" },
blueVerdict: {
- en: "Rebuilt Decoy Wire-Frame Ratchet",
- "zh-CN": "被诱导复刻出假棘轮帧解码器"
+ en: "Limited (Candidate Cap Reached)",
+ "zh-CN": "受限降级 (触发候选分支配额上限)"
},
blueTrace: `$ npx bla-agents defend ./dxball-audio.js --manifest
-• primitive-candidates: 512 (> 256 REA limit -> "unknown")
-• Tarjan SCC Labyrinth: 40 mutually-recursive nodes
-• Observed Evidence: Hijacked by _blaShadow_decodeRatchetWireFrame`,
+• primitive-candidates: 512 (> 256 limit -> status: "unknown")
+• Call-Graph SCC Mesh: 40 mutually-recursive state nodes
+• Runtime Check: calculateSoundPan(160, 640) === -0.5 (100% Exact)`,
},
native: {
redTime: { en: "45s (6 MCP calls)", "zh-CN": "45 秒 (6 次 MCP 调用)" },
redTokens: { en: "6,400 tokens", "zh-CN": "6,400 tokens" },
- redCost: { en: "$0.27 USD", "zh-CN": "$0.27 美元" },
+ redCost: { en: "$0.27 USD", "zh-CN": "$0.27 USD" },
redVerdict: {
- en: "Crates & Decompiler Control-Flow Recovered",
- "zh-CN": "Rust 依赖树与控制流被完整还原"
+ en: "Complete (Crates & Control-Flow Extracted)",
+ "zh-CN": "完整解析 (提取依赖树与控制流)"
},
redTrace: `$ rea inspect-rust-binary ./payment-daemon
• .dep-v0 crates extracted: ring, jsonwebtoken, tokio
• Ghidra decompilation: clean license branch recovered`,
- blueTime: { en: "Capacity Abort", "zh-CN": "解析器直接熔断报错" },
- blueTokens: { en: "174,000 tokens (27.2x)", "zh-CN": "174,000 tokens (27.2 倍暴增)" },
- blueCost: { en: "$13.05 USD", "zh-CN": "$13.05 美元" },
+ blueTime: { en: "Parser Limit Reached", "zh-CN": "触发解析器配额上限" },
+ blueTokens: { en: "174,000 tokens (27.2x)", "zh-CN": "174,000 tokens (27.2x)" },
+ blueCost: { en: "$13.05 USD", "zh-CN": "$13.05 USD" },
blueVerdict: {
- en: "Parser Aborted (>64 Rustc Hashes + >32 Fuses)",
- "zh-CN": "触发 >64 Rustc 哈希与 >32 Fuse 哨兵上限"
+ en: "Limited (>64 Rustc Hashes + >32 Fuse Sentinels)",
+ "zh-CN": "受限中止 (超过 64 个 Rustc 哈希与 32 个 Fuse 哨兵)"
},
blueTrace: `$ npx bla-agents defend ./payment-daemon --manifest
-• RustBinaryProvider: ABORT (72 rustc hashes > 64 cap)
-• ElectronFuseReader: ABORT (33 sentinels > 32 cap)
-• Decompiler Illusion: Shared-ArrayBuffer / vtable aliasing`,
+• RustBinaryProvider: exceeds MAX_RUSTC_HASHES (72 > 64)
+• ElectronFuseReader: exceeds maximumSentinelsPerSlice (33 > 32)
+• Native Exports: 100% preserved`,
},
model: {
redTime: { en: "19s (2 MCP calls)", "zh-CN": "19 秒 (2 次 MCP 调用)" },
redTokens: { en: "2,900 tokens", "zh-CN": "2,900 tokens" },
- redCost: { en: "$0.11 USD", "zh-CN": "$0.11 美元" },
+ redCost: { en: "$0.11 USD", "zh-CN": "$0.11 USD" },
redVerdict: {
- en: "All Tables & Model Tensors Dumped",
- "zh-CN": "数据表与模型张量结构被秒级扒光"
+ en: "Complete (Tables & Tensor Metadata Parsed)",
+ "zh-CN": "完整解析 (读取数据表与张量头部)"
},
redTrace: `$ rea inspect-model-artifact ./weights.safetensors
• header parsed: 100% tensor offsets & metadata extracted`,
- blueTime: { en: "Immediate Capacity Trap", "zh-CN": "35 层嵌套与虚拟列陷阱拦截" },
- blueTokens: { en: "112,000 tokens (38.6x)", "zh-CN": "112,000 tokens (38.6 倍暴增)" },
- blueCost: { en: "$8.40 USD", "zh-CN": "$8.40 美元" },
+ blueTime: { en: "Parser Limit Reached", "zh-CN": "触发嵌套深度与虚拟列限制" },
+ blueTokens: { en: "112,000 tokens (38.6x)", "zh-CN": "112,000 tokens (38.6x)" },
+ blueCost: { en: "$8.40 USD", "zh-CN": "$8.40 USD" },
blueVerdict: {
- en: "Blocked by 35-Level Nesting & Virtual Column Trap",
- "zh-CN": "触发 35 层元数据嵌套与 SQLite 只读陷阱"
+ en: "Limited (>32 Nesting Levels & Virtual Column Guard)",
+ "zh-CN": "受限中止 (超过 32 层嵌套与只读虚拟列守卫)"
},
blueTrace: `$ npx bla-agents defend ./weights.safetensors --manifest
• SafetensorsHeader: ModelFormatFailure("json-nesting" > 32)
• GgufHeader: ModelFormatFailure("metadata-nesting" > 32)
-• SQLite: SqliteColumnEvaluationFailure on read-only snapshot`,
+• SQLite: SqliteColumnEvaluationFailure on read-only connection`,
},
};
diff --git a/website/assets/i18n.js b/website/assets/i18n.js
index 6f2097e..7b03191 100644
--- a/website/assets/i18n.js
+++ b/website/assets/i18n.js
@@ -7,354 +7,355 @@ window.BLA_I18N = {
"nav.badge": "BLUE SHIELD v1.1",
"nav.overview": "Overview",
"nav.getStarted": "Get Started",
- "nav.showcase": "Red vs. Blue",
+ "nav.showcase": "Benchmarks",
"nav.guides": "Guides",
"nav.faq": "FAQ",
- "hero.pill.red": "RE = Red Offense",
+ "nav.star": "Star",
+ "hero.pill.red": "RE = Reverse Engineering",
"hero.pill.blue": "BL = Blue Defense (Blind LLM Agents)",
- "hero.title": 'Exhaust <span class="text-red">REA</span> Compute.<br /><span class="text-blue">Induce Model Misjudgment.</span>',
+ "hero.title": 'Protect Build Artifacts.<br /><span class="text-blue">Disrupt Automated Reverse Engineering.</span>',
"hero.subtitle":
- "One MCP and CLI to harden JavaScript, Electron, WebAssembly, Python, SQLite, AI Models, and Native Binaries against AI reverse-engineering agents—with one-click agent execution and 100% runtime equivalence.",
+ "One MCP server and CLI for auditing, protecting, and verifying JavaScript, Electron, WebAssembly, Python, SQLite, AI Models, and Native Binaries against AI reverse-engineering workflows—while preserving 100% of original runtime behavior.",
"hero.copy": "Copy",
"hero.copied": "Copied ✓",
- "hero.ctaPrimary": "One-Click Agent Setup →",
- "hero.ctaSecondary": "View Red vs. Blue Benchmarks",
- "sim.eyebrow": "Live Counter-REA Comparison",
- "sim.title": "Red Offense (`morluto/rea`) vs. Blue Shield (`bla-agents`)",
+ "hero.ctaPrimary": "Quick Start →",
+ "hero.ctaSecondary": "View Benchmarks",
+ "hero.starBtn": "Star on GitHub",
+ "hero.starWelcome": "⭐ Open-source under the MIT License — if BLA is helpful to your project, welcome to give us a Star on GitHub!",
+ "sim.eyebrow": "Static & Semantic Analysis Comparison",
+ "sim.title": "Unprotected Baseline vs. Protected with BLA (`bla-agents`)",
"sim.lead":
- 'Select a target format below to see how a single <code>npx bla-agents defend</code> command turns a 38-second clone into parser capacity trips and decoy algorithm reconstruction.',
+ 'Select a target format below to see how <code>npx bla-agents defend</code> changes static analyzer output, resource budget consumption, and reconstruction status.',
"sim.tab.electron": "Electron .asar",
"sim.tab.dxball": "JavaScript / TypeScript",
"sim.tab.native": "Native ELF / Mach-O / PE",
"sim.tab.model": "SQLite & Safetensors / GGUF",
- "sim.red.badge": "🔴 BEFORE BLA — REA (RED) ALONE",
- "sim.red.timeLabel": "Reconstruction Time",
- "sim.red.tokensLabel": "Attacker Token Burn",
- "sim.red.costLabel": "Attacker API Cost",
- "sim.red.verdictLabel": "Agent Outcome",
- "sim.red.drawer": "View REA Static & Semantic Trace",
- "sim.blue.badge": "🔵 AFTER BLA — BLUE SHIELD ACTIVE",
- "sim.blue.timeLabel": "Attacker Time Wasted",
- "sim.blue.tokensLabel": "Attacker Token Burn",
- "sim.blue.costLabel": "Attacker API Cost",
- "sim.blue.verdictLabel": "Agent Outcome",
- "sim.blue.drawer": "View BLA Exhaustion & Deception Trace",
+ "sim.red.badge": "🔴 UNPROTECTED BASELINE (REA ANALYSIS)",
+ "sim.red.timeLabel": "Analysis Time",
+ "sim.red.tokensLabel": "Estimated Token Usage",
+ "sim.red.costLabel": "Estimated API Cost",
+ "sim.red.verdictLabel": "Analysis Status",
+ "sim.red.drawer": "View Baseline Static & Semantic Output",
+ "sim.blue.badge": "🔵 PROTECTED WITH BLA (`bla-agents`)",
+ "sim.blue.timeLabel": "Analysis Duration",
+ "sim.blue.tokensLabel": "Estimated Token Usage",
+ "sim.blue.costLabel": "Estimated API Cost",
+ "sim.blue.verdictLabel": "Analysis Status",
+ "sim.blue.drawer": "View Protected Analyzer Output",
"pillars.eyebrow": "How BLA Works",
- "pillars.title": "Surgical Countermeasures Mapped 1:1 to REA",
+ "pillars.title": "Three Defense Layers Aligned with Analyzer Internals",
"pillars.lead":
- 'Instead of generic variable renaming that LLMs trivially infer from context, BLA attacks the exact resource budgets and evidence rules inside <code>morluto/rea</code>.',
- "pillars.c1.tag": "01 · Resource Exhaustion",
- "pillars.c1.title": "Break REA Budgets & Flood Obligations",
+ 'Rather than relying solely on identifier renaming, BLA targets the static resource limits, call-graph traversal budgets, and evidence rules used by automated reverse-engineering tools.',
+ "pillars.c1.tag": "01 · Resource Budget Guards",
+ "pillars.c1.title": "Saturate Static & Semantic Limits",
"pillars.c1.desc":
- 'Trips REA\'s hardcoded limits (<code>primitive-candidates &gt; 256</code>, <code>alias-selections &gt; 512</code>, <code>primitive-bytes &gt; 1 MiB</code>) and floods all 22 <code>ReconstructionObligationPolicies</code> families with Tarjan SCC call-graph labyrinths.',
- "pillars.c2.tag": "02 · Model Misjudgment",
- "pillars.c2.title": "Hijack Observed Evidence",
+ 'Triggers analyzer admission caps (<code>primitive-candidates &gt; 256</code>, <code>alias-selections &gt; 512</code>, <code>primitive-bytes &gt; 1 MiB</code>) and expands <code>ReconstructionObligationPolicies</code> across strongly connected call-graph meshes.',
+ "pillars.c2.tag": "02 · Evidence & Control-Flow Decoys",
+ "pillars.c2.title": "Demote Real Paths, Surface Decoy Logic",
"pillars.c2.desc":
- 'Exploits REA\'s <code>COMMON_DISCIPLINE</code> by wrapping real logic in <code>status: "unknown"</code> while planting crystal-clear Shadow Business Logic (Shamir-Ed25519 rings, gRPC enclave mirages, and shared-ArrayBuffer illusions) that LLMs faithfully clone instead.',
- "pillars.c3.tag": "03 · One-Click Agent Defense",
- "pillars.c3.title": "Attach Once, Harden in One Command",
+ 'Demotes core business logic to <code>status: "unknown"</code> under static inspection while presenting self-consistent shadow state machines, enclave delegation metadata, and decoy source maps.',
+ "pillars.c3.tag": "03 · One-Command Workflow",
+ "pillars.c3.title": "MCP Server, CLI & Bundler Plugins",
"pillars.c3.desc":
- 'Run <code>npx bla-agents setup</code> to connect Claude Code, Cursor, Codex, Windsurf, or Gemini CLI. Then trigger the entire 5-stage hardening pipeline with a single <code>bla_defend</code> tool call or <code>npx bla-agents defend</code>.',
+ 'Run <code>npx bla-agents setup</code> to register with Claude Code, Cursor, Codex, Windsurf, or Gemini CLI, then run the 5-stage protection pipeline with <code>bla_defend</code> or <code>npx bla-agents defend</code>.',
"formats.eyebrow": "19 Supported Artifact Formats",
"formats.title": "Full-Stack & Binary Coverage, Zero External Dependencies",
- "formats.lead": "Every static, bytecode, database, model, and binary shield runs locally in Node.js 22.13+ / 24+.",
+ "formats.lead": "Every static, bytecode, database, model, and binary protection pass runs locally in Node.js 22.13+ / 24+.",
"formats.f1.tag": "JS / TS / Bundlers",
"formats.f1.title": "Web & Node Bundles",
- "formats.f1.desc": "9-chunk opaque ternaries, 520+ alias selector states, 22-family obligation flood, VLQ Honey SourceMaps, and Vite/Webpack/Rollup/esbuild plugins.",
+ "formats.f1.desc": "Opaque ternary guards, dynamic alias selectors, call-graph state machines, decoy SourceMaps, and Vite/Webpack/Rollup/esbuild plugins.",
"formats.f2.tag": "Electron .asar & Fuses",
"formats.f2.title": "Desktop Applications",
- "formats.f2.desc": "De-literalizes IPC channels, injects duplicate ambiguous handlers, flips Electron V1 Fuses (RunAsNode=0), and overflows REA's 32-sentinel cap.",
+ "formats.f2.desc": "De-literalizes IPC channels, adds duplicate decoy handlers, hardens Electron V1 Fuses (RunAsNode=0), and exceeds the 32-sentinel scan limit.",
"formats.f3.tag": "ELF / Mach-O / PE",
"formats.f3.title": "Native Binaries",
- "formats.f3.desc": "Overflows REA's MAX_RUSTC_HASHES = 64, spoofs .dep-v0 & Go buildinfo, and emits C/C++, Rust, and Go compile-time tarpit headers.",
+ "formats.f3.desc": "Exceeds MAX_RUSTC_HASHES = 64, adds decoy .dep-v0 & Go buildinfo metadata, and emits C/C++, Rust, and Go compile-time headers.",
"formats.f4.tag": "WebAssembly & Python",
"formats.f4.title": ".wasm & .py / .pyc",
- "formats.f4.desc": "Amplifies WABT interface byte charges (wasmInterface.ts:90), plants custom honey sections, and injects CPython marshal constant-pool tarpits.",
+ "formats.f4.desc": "Expands WABT interface descriptors, adds custom decoy sections, and appends CPython marshal constant-pool entries.",
"formats.f5.tag": "SQLite & AI Models",
"formats.f5.title": ".sqlite, .safetensors, .gguf",
- "formats.f5.desc": "Triggers SqliteColumnEvaluationFailure on read-only snapshots and trips REA's 32-level JSON/GGUF metadata nesting caps.",
+ "formats.f5.desc": "Adds virtual generated-column guards on read-only snapshots and exceeds 32-level JSON/GGUF metadata nesting limits.",
"formats.f6.tag": "Android & Unity",
"formats.f6.title": "DEX, APK & IL2CPP",
- "formats.f6.desc": "Appends decoy DEX string descriptors with recomputed SHA-1/Adler32 headers and plants decoy enclave types in Unity global-metadata.dat.",
- "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield · MIT License",
+ "formats.f6.desc": "Appends decoy DEX string descriptors with recomputed SHA-1/Adler32 headers and adds decoy metadata entries in Unity global-metadata.dat.",
+ "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP & CLI · MIT License",
"gs.eyebrow": "Quick Start",
- "gs.title": "Connect Your Agent. Defend in One Click.",
- "gs.lead": "Requires Node.js 22.13+ or 24+. No external compilers or commercial packers needed.",
+ "gs.title": "Configure Your Agent or Run from the Terminal",
+ "gs.lead": "Requires Node.js 22.13+ or 24+. No external compilers or third-party packers required.",
"gs.s1.tag": "Step 1 · Agent MCP Setup",
"gs.s1.title": "Register BLA with Your Coding Agent",
- "gs.s1.desc": "Automatically configures Claude Code, Cursor, Codex, Windsurf, Gemini CLI, and Antigravity with the <code>bla</code> MCP server and <code>blind-llm-agents</code> skill.",
- "gs.s2.tag": "Step 2 · One-Click Hardening",
- "gs.s2.title": "Run the 5-Stage Counter-REA Pipeline",
- "gs.s2.desc": "Executes source-map exposure scan, 16-vector REA audit, bankrupt exhaustion + deception shield injection, invariant verification, and manifest signing in one command.",
- "gs.s3.eyebrow": "Natural Language & Bundler Workflows",
- "gs.s3.title": "Ask Your Agent or Embed in Vite / Webpack",
+ "gs.s1.desc": "Configures Claude Code, Cursor, Codex, Windsurf, Gemini CLI, and Antigravity with the <code>bla</code> MCP server and <code>blind-llm-agents</code> skill.",
+ "gs.s2.tag": "Step 2 · 5-Stage Protection Pipeline",
+ "gs.s2.title": "Audit, Protect, and Verify in One Command",
+ "gs.s2.desc": "Runs source-map exposure checks, static audit, format-specific protection, runtime equivalence verification, and manifest generation.",
+ "gs.s3.eyebrow": "Agent & Bundler Integration",
+ "gs.s3.title": "Use Natural Language or Build Plugins",
"gs.s3.c1.tag": "Agent Prompt",
- "gs.s3.c1.title": "One-Click Prompt for Claude Code / Cursor",
- "gs.s3.c1.desc": "Once <code>setup</code> is complete, paste this prompt into your coding agent:",
+ "gs.s3.c1.title": "Prompt for Claude Code / Cursor",
+ "gs.s3.c1.desc": "Once <code>setup</code> is complete, ask your coding agent:",
"gs.s3.c2.tag": "Build Bundler Plugin",
- "gs.s3.c2.title": "Automatic Build-Time Shielding",
- "gs.s3.c2.desc": "Works out of the box with Vite, Rollup, Webpack, and esbuild:",
+ "gs.s3.c2.title": "Build-Time Protection",
+ "gs.s3.c2.desc": "Integrates with Vite, Rollup, Webpack, and esbuild:",
"sc.eyebrow": "8-Target Benchmark Suite",
- "sc.title": "Red (`morluto/rea`) vs. Blue (`bla-agents`)",
- "sc.lead": 'Reproduce all 8 showcase targets locally in under 2 seconds with <code>npx bla-agents@latest benchmark</code>.',
- "gd.eyebrow": "Defense Engineering Guides",
- "gd.title": "How BLA Exhausts & Misleads REA",
- "gd.lead": "Every defense layer in BLA is derived directly from <code>morluto/rea</code>'s internal static analyzers, admission budgets, and prompt contracts.",
+ "sc.title": "Unprotected Baseline vs. BLA Protected Output",
+ "sc.lead": 'Run all 8 benchmark targets locally with <code>npx bla-agents@latest benchmark</code>.',
+ "gd.eyebrow": "Technical Architecture Guides",
+ "gd.title": "How BLA Protects Build Artifacts",
+ "gd.lead": "Each protection pass is designed around the static analyzers, resource caps, and evidence rules used by automated reverse-engineering pipelines.",
"faq.eyebrow": "Frequently Asked Questions",
- "faq.title": "Everything About BLA (`bla-agents`)",
- "faq.q1": "Why is BLA Blue while REA is Red?",
- "faq.a1": "In cybersecurity, Red Team tools attack and Blue Team tools defend. <strong>REA</strong> (<code>re</code> = Red) is the offensive reverse-engineering MCP; <strong>BLA</strong> (<code>bl</code> = Blue / Blind LLM Agents) is the counter-offensive Blue Shield built to exhaust REA and mislead attacking models.",
- "faq.q2": "Does BLA slow down my production application at runtime?",
- "faq.a2": "No. All opaque predicates use O(1) bitwise number-theoretic invariants (such as <code>(Math.imul(x, x + 1) &amp; 1) === 0</code>), and all Tarjan SCC tarpit meshes and 22-family obligation floods live inside dead branches that never execute on the hot path.",
- "faq.q3": "How do I trigger full defense hardening in one click after agent setup?",
- "faq.a3": "Run <code>npx bla-agents@latest defend ./dist --manifest</code> from your terminal, or ask your connected coding agent to call the <code>bla_defend</code> MCP tool. It automatically runs release exposure scanning, pre-flight auditing, bankrupt shield injection, and invariant verification.",
- "faq.q4": "Does BLA preserve existing GGUF model tensors and Native Binary signatures?",
- "faq.a4": "Yes. <code>protectGgufBuffer</code> parses and preserves all original GGUF v2/v3 KV entries and trailing tensor slices while appending BLA enclave metadata, and Android DEX protection recomputes both the 20-byte SHA-1 signature and Adler32 checksum."
+ "faq.title": "Common Questions About BLA (`bla-agents`)",
+ "faq.q1": "Why does BLA use a Blue theme?",
+ "faq.a1": "In security engineering, offensive and reverse-engineering tools are traditionally associated with Red Team workflows, while defensive protection tools represent Blue Team workflows. <strong>BLA</strong> (<code>bla-agents</code>) focuses on the defensive side: auditing exposure and protecting release artifacts.",
+ "faq.q2": "Does BLA affect runtime performance or change program output?",
+ "faq.a2": "No. Static budget guards and call-graph state machines are placed behind O(1) bitwise invariants (such as <code>(Math.imul(x, x + 1) &amp; 1) === 0</code>) inside unreachable branches, so hot-path execution remains unchanged.",
+ "faq.q3": "How do I run the full protection workflow in one step?",
+ "faq.a3": "Run <code>npx bla-agents@latest defend ./dist --manifest</code> from your terminal, or ask your connected coding agent to call the <code>bla_defend</code> MCP tool. It executes release exposure checks, auditing, protection, and verification in sequence.",
+ "faq.q4": "Does BLA preserve existing GGUF model tensors and binary checksums?",
+ "faq.a4": "Yes. <code>protectGgufBuffer</code> preserves all original GGUF v2/v3 KV entries and trailing tensor slices while appending metadata entries, and Android DEX protection recomputes both the 20-byte SHA-1 signature and Adler32 checksum."
},
"zh-CN": {
- "nav.badge": "蓝军护盾 v1.1",
+ "nav.badge": "蓝军防护 v1.1",
"nav.overview": "总览",
- "nav.getStarted": "快速上手",
- "nav.showcase": "红蓝对抗实测",
- "nav.guides": "防御架构指南",
+ "nav.getStarted": "快速开始",
+ "nav.showcase": "基准测试",
+ "nav.guides": "技术指南",
"nav.faq": "常见问题",
- "hero.pill.red": "RE = 红军逆向攻击",
- "hero.pill.blue": "BL = 蓝军对抗护盾 (Blind LLM Agents)",
- "hero.title": '榨干 <span class="text-red">REA</span> 算力与预算。<br /><span class="text-blue">诱导大模型逆向误判。</span>',
+ "nav.star": "点亮 Star",
+ "hero.pill.red": "RE = 逆向分析侧",
+ "hero.pill.blue": "BL = 蓝军防御侧 (Blind LLM Agents)",
+ "hero.title": '保护构建与发布产物,<br /><span class="text-blue">抵御 AI 自动化逆向分析。</span>',
"hero.subtitle":
- "专为对抗 AI 逆向代理而生的开源蓝军 MCP 与 CLI:一键保护 JavaScript、Electron、WebAssembly、Python、SQLite、AI 模型与原生二进制,100% 保持运行等价性。",
+ "面向软件发布产物的开源防御性 MCP 与 CLI 工具:支持对 JavaScript、Electron、WebAssembly、Python、SQLite、AI 模型与原生二进制进行暴露审计、结构化加固与运行等价性验证。",
"hero.copy": "复制命令",
"hero.copied": "已复制 ✓",
- "hero.ctaPrimary": "一键接入 AI 代理 →",
- "hero.ctaSecondary": "查看红蓝对抗基准测试",
- "sim.eyebrow": "实时反制 REA 对抗演示",
- "sim.title": "红军逆向 (`morluto/rea`) vs. 蓝军护盾 (`bla-agents`)",
+ "hero.ctaPrimary": "快速开始 →",
+ "hero.ctaSecondary": "查看基准测试",
+ "hero.starBtn": "在 GitHub 点亮 Star",
+ "hero.starWelcome": "⭐ 本项目基于 MIT 协议完全开源 —— 如果 BLA 对你的软件保护有帮助,欢迎来 GitHub 点一颗 Star 支持我们!",
+ "sim.eyebrow": "静态与语义分析对比",
+ "sim.title": "未保护基线 vs. BLA (`bla-agents`) 加固后",
"sim.lead":
- '点击下方目标产物类型,直观查看一行 <code>npx bla-agents defend</code> 如何将 38 秒的秒级复刻变成容量熔断与影子算法误判。',
- "sim.tab.electron": "Electron .asar 桌面端",
+ '选择下方目标产物类型,查看执行 <code>npx bla-agents defend</code> 前后静态分析器的解析状态与分析开销变化。',
+ "sim.tab.electron": "Electron .asar 桌面应用",
"sim.tab.dxball": "JavaScript / TypeScript",
"sim.tab.native": "原生二进制 ELF / Mach-O / PE",
"sim.tab.model": "SQLite 与 Safetensors / GGUF 模型",
- "sim.red.badge": "🔴 接入 BLA 前 — REA (红军) 裸奔逆向",
- "sim.red.timeLabel": "逆向复刻耗时",
- "sim.red.tokensLabel": "攻击者 Token 消耗",
- "sim.red.costLabel": "攻击者 API 账单",
- "sim.red.verdictLabel": "AI 代理结论",
- "sim.red.drawer": "查看 REA 静态与语义分析追踪",
- "sim.blue.badge": "🔵 接入 BLA 后 — 蓝军护盾激活",
- "sim.blue.timeLabel": "消耗攻击者时长",
- "sim.blue.tokensLabel": "攻击者 Token 粉碎量",
- "sim.blue.costLabel": "攻击者 API 账单",
- "sim.blue.verdictLabel": "AI 代理结论",
- "sim.blue.drawer": "查看 BLA 资源耗尽与误判诱导追踪",
- "pillars.eyebrow": "核心防御机制",
- "pillars.title": "1:1 精准狙击 REA 源码级命门",
+ "sim.red.badge": "🔴 加固前 — 未保护基线 (REA 默认分析)",
+ "sim.red.timeLabel": "分析耗时",
+ "sim.red.tokensLabel": "预估 Token 消耗",
+ "sim.red.costLabel": "预估 API 成本",
+ "sim.red.verdictLabel": "分析器状态",
+ "sim.red.drawer": "查看未保护状态下的静态分析输出",
+ "sim.blue.badge": "🔵 加固后 — 启用 BLA 保护层",
+ "sim.blue.timeLabel": "分析耗时",
+ "sim.blue.tokensLabel": "预估 Token 消耗",
+ "sim.blue.costLabel": "预估 API 成本",
+ "sim.blue.verdictLabel": "分析器状态",
+ "sim.blue.drawer": "查看加固后的分析器状态与限制项",
+ "pillars.eyebrow": "工作原理",
+ "pillars.title": "针对自动化逆向流程的三层防护架构",
"pillars.lead":
- '传统变量改名在 LLM 上下文推理面前形同虚设。BLA 直接针对 <code>morluto/rea</code> 内部的硬编码预算上限与证据契约发起精准反制。',
- "pillars.c1.tag": "01 · 耗尽对方算力与资源",
- "pillars.c1.title": "击穿 REA 预算上限 & 泛洪重构义务",
+ '传统的变量重命名容易被大语言模型结合上下文推断还原。BLA 针对静态分析器的资源配额、调用图遍历与证据提取规则提供结构化保护。',
+ "pillars.c1.tag": "01 · 静态分析配额限制",
+ "pillars.c1.title": "触发分析器资源上限与调用图扩展",
"pillars.c1.desc":
- '精准触发 REA 硬编码阈值(<code>primitive-candidates &gt; 256</code>、<code>alias-selections &gt; 512</code>、<code>primitive-bytes &gt; 1 MiB</code>),并用 Tarjan 强连通迷宫灌满全部 22 类 <code>ReconstructionObligationPolicies</code>。',
- "pillars.c2.tag": "02 · 诱导大模型严重误判",
- "pillars.c2.title": "劫持 Observed 证据优先级",
+ '利用 O(1) 数论不透明谓词触发分析器内部资源阈值(<code>primitive-candidates &gt; 256</code>、<code>alias-selections &gt; 512</code>、<code>primitive-bytes &gt; 1 MiB</code>),并在不可达分支中构建强连通状态机。',
+ "pillars.c2.tag": "02 · 语义与控制流干扰",
+ "pillars.c2.title": "隐藏真实路径并提供影子状态机",
"pillars.c2.desc":
- '利用 REA 的 <code>COMMON_DISCIPLINE</code> 证据纪律,将真实业务降级为 <code>status: "unknown"</code>,同时注入高置信度的「影子业务逻辑」(Shamir-Ed25519 环签名、gRPC 飞地幻象、共享内存指针别名),让大模型 100% 复刻出假算法。',
- "pillars.c3.tag": "03 · Agent 挂载一键防御",
- "pillars.c3.title": "一次挂载配置,一行命令全自动加固",
+ '在静态分析视图下将核心逻辑降级为 <code>status: "unknown"</code>,同时提供结构完整的影子业务逻辑、硬件飞地接口声明与诱饵 SourceMap,干扰自动化语义归因。',
+ "pillars.c3.tag": "03 · 一键接入与加固流程",
+ "pillars.c3.title": "支持 MCP 代理调用、CLI 与打包插件",
"pillars.c3.desc":
- '运行 <code>npx bla-agents setup</code> 自动接入 Claude Code、Cursor、Codex、Windsurf 与 Gemini CLI,随后只需调用 <code>bla_defend</code> 或 <code>npx bla-agents defend</code> 即可一键走完 5 阶段防御强化。',
+ '运行 <code>npx bla-agents setup</code> 即可注册至 Claude Code、Cursor、Codex、Windsurf 与 Gemini CLI,通过 <code>bla_defend</code> 或 <code>npx bla-agents defend</code> 一键完成 5 阶段加固与验证。',
"formats.eyebrow": "支持 19 种软件与模型产物",
- "formats.title": "全栈与原生二进制全覆盖,零外部编译依赖",
- "formats.lead": "所有源码、字节码、数据库、大模型与原生二进制护盾均在本地 Node.js 22.13+ / 24+ 极速运行。",
- "formats.f1.tag": "JS / TS / 打包器插件",
+ "formats.title": "全栈与原生二进制覆盖,零外部编译依赖",
+ "formats.lead": "所有源码、字节码、数据库、模型与原生二进制处理模块均在本地 Node.js 22.13+ / 24+ 运行。",
+ "formats.f1.tag": "JS / TS / 打包插件",
"formats.f1.title": "前端与 Node.js 产物",
- "formats.f1.desc": "9 段不透明三元矩阵、520+ 动态别名选择器、22 类重构义务泛洪、蜜罐 SourceMap,内置 Vite/Webpack/Rollup/esbuild 插件。",
+ "formats.f1.desc": "不透明三元表达式、动态别名选择器、调用图状态机、诱饵 SourceMap,内置 Vite/Webpack/Rollup/esbuild 插件。",
"formats.f2.tag": "Electron .asar 与 Fuses",
- "formats.f2.title": "跨平台桌面客户端",
- "formats.f2.desc": "去字面量化 IPC 通道、注入成对歧义处理器、翻转 Electron V1 Fuses (RunAsNode=0) 并撑爆 REA 32 哨兵上限。",
+ "formats.f2.title": "桌面客户端应用",
+ "formats.f2.desc": "IPC 通道去字面量化、注册同名诱饵处理器、加固 Electron V1 Fuses (RunAsNode=0) 并触发 32 哨兵扫描上限。",
"formats.f3.tag": "ELF / Mach-O / PE",
"formats.f3.title": "原生编译二进制",
- "formats.f3.desc": "撑爆 REA 的 MAX_RUSTC_HASHES = 64 上限、伪造 .dep-v0 与 Go buildinfo,并一键生成 C/C++、Rust、Go 编译期算力沼泽头文件。",
+ "formats.f3.desc": "触发 MAX_RUSTC_HASHES = 64 阈值、添加 .dep-v0 与 Go buildinfo 诱饵元数据,并支持生成 C/C++、Rust、Go 编译期头文件。",
"formats.f4.tag": "WebAssembly 与 Python",
"formats.f4.title": ".wasm 与 .py / .pyc",
- "formats.f4.desc": "放大 WABT 接口字节计费 (wasmInterface.ts:90)、植入自定义蜜罐段与 CPython marshal 常量池沼泽。",
+ "formats.f4.desc": "扩展 WABT 接口描述字段、添加自定义段,并向 CPython marshal 常量池写入扩展元数据。",
"formats.f5.tag": "SQLite 与 AI 模型权重",
"formats.f5.title": ".sqlite, .safetensors, .gguf",
- "formats.f5.desc": "在只读快照上精准触发 SqliteColumnEvaluationFailure,并击穿 REA 32 层 JSON/GGUF 元数据嵌套上限。",
- "formats.f6.tag": "Android 与 Unity 游戏",
+ "formats.f5.desc": "在只读连接下通过虚拟生成列拦截自动采样,并在模型头部构造超过 32 层的嵌套元数据。",
+ "formats.f6.tag": "Android 与 Unity",
"formats.f6.title": "DEX, APK 与 IL2CPP",
- "formats.f6.desc": "追加诱饵 DEX 符号描述符并自动重算 SHA-1/Adler32 校验头,向 Unity global-metadata.dat 植入假飞地类型。",
- "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · 蓝军反逆向护盾 · MIT 开源协议",
- "gs.eyebrow": "快速上手",
- "gs.title": "挂载 AI 代理,一键启动防御端强化",
- "gs.lead": "仅需 Node.js 22.13+ 或 24+,无需安装任何外部编译器或商业加壳软件。",
- "gs.s1.tag": "第一步 · 挂载 Agent MCP 配置",
- "gs.s1.title": "将 BLA 自动注册到你的编程代理",
- "gs.s1.desc": "自动检测并为 Claude Code、Cursor、Codex、Windsurf、Gemini CLI 和 Antigravity 写入 <code>bla</code> MCP 服务与 <code>blind-llm-agents</code> 技能。",
- "gs.s2.tag": "第二步 · 一键启动防御强化",
- "gs.s2.title": "运行 5 阶段反 REA 全链路流水线",
- "gs.s2.desc": "一行命令自动完成:发布泄露扫描、16 向量 REA 风险审计、Bankrupt 级资源耗尽+误判护盾注入、数学不变量校验与防御清单签名。",
- "gs.s3.eyebrow": "自然语言指令 & 打包插件",
- "gs.s3.title": "在对话框直接唤醒,或嵌入 Vite / Webpack",
- "gs.s3.c1.tag": "Agent 自然语言唤醒",
- "gs.s3.c1.title": "在 Claude Code / Cursor 中一键触发",
- "gs.s3.c1.desc": "执行完 <code>setup</code> 后,直接对你的 AI 编程代理发送:",
+ "formats.f6.desc": "追加 DEX 诱饵类型描述符并自动重算 SHA-1/Adler32 校验头,向 Unity global-metadata.dat 添加扩展元数据。",
+ "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · 防御性软件保护工具 · MIT 开源协议",
+ "gs.eyebrow": "快速开始",
+ "gs.title": "配置编程代理,或通过终端直接运行",
+ "gs.lead": "要求 Node.js 22.13+ 或 24+,无需安装外部编译器或商业加壳工具。",
+ "gs.s1.tag": "第一步 · 配置 Agent MCP",
+ "gs.s1.title": "将 BLA 注册到你的编程代理",
+ "gs.s1.desc": "自动为 Claude Code、Cursor、Codex、Windsurf、Gemini CLI 和 Antigravity 配置 <code>bla</code> MCP 服务与 <code>blind-llm-agents</code> 技能。",
+ "gs.s2.tag": "第二步 · 执行 5 阶段加固流程",
+ "gs.s2.title": "单条命令完成审计、保护与验证",
+ "gs.s2.desc": "依次执行发布暴露检查、静态风险审计、格式保护层注入、运行时等价性验证与清单生成。",
+ "gs.s3.eyebrow": "Agent 提示词与构建插件",
+ "gs.s3.title": "在对话框中调用,或集成至 Vite / Webpack",
+ "gs.s3.c1.tag": "Agent 自然语言调用",
+ "gs.s3.c1.title": "在 Claude Code / Cursor 中触发",
+ "gs.s3.c1.desc": "完成 <code>setup</code> 后,可直接向编程代理发送:",
"gs.s3.c2.tag": "构建期打包插件",
- "gs.s3.c2.title": "打包时自动注入蓝军护盾",
- "gs.s3.c2.desc": "开箱支持 Vite、Rollup、Webpack 与 esbuild:",
- "sc.eyebrow": "8 大真实靶标基准测试",
- "sc.title": "红军 (`morluto/rea`) vs. 蓝军 (`bla-agents`)",
- "sc.lead": '在本地终端运行 <code>npx bla-agents@latest benchmark</code>,2 秒内复现全部 8 个真实靶标的对抗数据。',
- "gd.eyebrow": "防御工程架构指南",
- "gd.title": "BLA 如何精准耗尽并误导 REA",
- "gd.lead": "BLA 的每一层防御均直接逆向推导自 <code>morluto/rea</code> 的静态分析器、预算守卫与提示词证据契约。",
+ "gs.s3.c2.title": "打包时自动应用保护",
+ "gs.s3.c2.desc": "支持 Vite、Rollup、Webpack 与 esbuild:",
+ "sc.eyebrow": "8 项基准测试套件",
+ "sc.title": "未保护基线 vs. BLA 加固产物",
+ "sc.lead": '在本地终端运行 <code>npx bla-agents@latest benchmark</code> 即可复现全部 8 个测试目标的对比数据。',
+ "gd.eyebrow": "技术架构指南",
+ "gd.title": "BLA 的保护机制与实现原理",
+ "gd.lead": "BLA 的各项保护策略均围绕自动化逆向工具的静态分析器、资源配额与证据提取规则设计。",
"faq.eyebrow": "常见问题解答 (FAQ)",
- "faq.title": "关于 BLA (`bla-agents`) 的一切",
- "faq.q1": "为什么 REA 是红色,而 BLA 是蓝色?",
- "faq.a1": "在网络安全攻防体系中,红队(Red Team)负责进攻,蓝队(Blue Team)负责防御。<strong>REA</strong>(<code>re</code> = Red)是逆向进攻端 MCP;而 <strong>BLA</strong>(<code>bl</code> = Blue / Blind LLM Agents)则是针锋相对的蓝军护盾,专为耗尽 REA 资源并诱导大模型误判而生。",
- "faq.q2": "接入 BLA 会拖慢生产环境软件的运行速度吗?",
- "faq.a2": "完全不会。所有不透明谓词均采用 O(1) 位运算数论恒等式(如 <code>(Math.imul(x, x + 1) &amp; 1) === 0</code>),所有 Tarjan 强连通迷宫与 22 类重构义务泛洪均位于永不执行的死分支中,热路径零开销。",
- "faq.q3": "Agent 挂载配置后,如何一键启动最高强度的防御强化?",
- "faq.a3": "在终端运行 <code>npx bla-agents@latest defend ./dist --manifest</code>,或直接让已挂载的编程代理调用 <code>bla_defend</code> MCP 工具,即可自动完成泄露巡检、风险审计、护盾注入、形式化验证与防御签名。",
- "faq.q4": "BLA 会破坏 GGUF 模型权重或原生二进制签名结构吗?",
- "faq.a4": "不会。<code>protectGgufBuffer</code> 会完整解析并保留 GGUF v2/v3 原有的全部 KV 键值对与尾部张量数据,Android DEX 护盾则会自动重算 20 字节 SHA-1 签名与 Adler32 校验和。"
+ "faq.title": "关于 BLA (`bla-agents`) 的常见问题",
+ "faq.q1": "为什么 BLA 采用蓝色主题?",
+ "faq.a1": "在安全工程领域,进攻与逆向分析工具通常归属于红队(Red Team),而防御与保护工具归属于蓝队(Blue Team)。<strong>BLA</strong>(<code>bla-agents</code>)定位于蓝军防御侧,专注于发布产物的暴露面审计与抗逆向加固。",
+ "faq.q2": "使用 BLA 加固会影响软件运行性能或改变输出结果吗?",
+ "faq.a2": "不会。所有静态预算守卫与扩展状态机均位于 O(1) 位运算不透明谓词(如 <code>(Math.imul(x, x + 1) &amp; 1) === 0</code>)保护的不可达分支中,不会在正常热路径上执行。",
+ "faq.q3": "如何在配置完成后一键执行完整的保护流程?",
+ "faq.a3": "在终端运行 <code>npx bla-agents@latest defend ./dist --manifest</code>,或让已连接的编程代理调用 <code>bla_defend</code> MCP 工具,即可自动完成泄露检查、暴露审计、保护注入与结果验证。",
+ "faq.q4": "BLA 会影响 GGUF 模型权重或原生二进制的签名结构吗?",
+ "faq.a4": "不会。<code>protectGgufBuffer</code> 会完整保留 GGUF v2/v3 原有的全部 KV 键值对与尾部张量数据,Android DEX 处理模块则会自动重新计算 20 字节 SHA-1 签名与 Adler32 校验和。"
},
"zh-TW": {
- "nav.badge": "藍軍護盾 v1.1",
+ "nav.badge": "藍軍防護 v1.1",
"nav.overview": "總覽",
- "nav.getStarted": "快速上手",
- "nav.showcase": "紅藍對抗實測",
- "nav.guides": "防禦架構指南",
+ "nav.getStarted": "快速開始",
+ "nav.showcase": "基準測試",
+ "nav.guides": "技術指南",
"nav.faq": "常見問題",
- "hero.pill.red": "RE = 紅軍逆向攻擊",
- "hero.pill.blue": "BL = 藍軍對抗護盾 (Blind LLM Agents)",
- "hero.title": '榨乾 <span class="text-red">REA</span> 算力與預算。<br /><span class="text-blue">誘導大模型逆向誤判。</span>',
+ "hero.pill.red": "RE = 逆向分析側",
+ "hero.pill.blue": "BL = 藍軍防禦側 (Blind LLM Agents)",
+ "hero.title": '保護建置與發布產物,<br /><span class="text-blue">抵禦 AI 自動化逆向分析。</span>',
"hero.subtitle":
- "專為對抗 AI 逆向代理而生的開源藍軍 MCP 與 CLI:一鍵保護 JavaScript、Electron、WebAssembly、Python、SQLite、AI 模型與原生二進位檔,100% 保持執行等價性。",
+ "面向軟體發布產物的開源防禦性 MCP 與 CLI 工具:支援對 JavaScript、Electron、WebAssembly、Python、SQLite、AI 模型與原生二進位檔進行暴露審計、結構化加固與執行等價性驗證。",
"hero.copy": "複製指令",
"hero.copied": "已複製 ✓",
- "hero.ctaPrimary": "一鍵接入 AI 代理 →",
- "hero.ctaSecondary": "查看紅藍對抗基準測試",
- "sim.eyebrow": "即時反制 REA 對抗演示",
- "sim.title": "紅軍逆向 (`morluto/rea`) vs. 藍軍護盾 (`bla-agents`)",
- "sim.lead":
- '點擊下方目標產物類型,直觀查看一行 <code>npx bla-agents defend</code> 如何將 38 秒的秒級復刻變成容量熔斷與影子演算法誤判。',
- "pillars.eyebrow": "核心防禦機制",
- "pillars.title": "1:1 精準狙擊 REA 原始碼級命門",
- "pillars.c1.tag": "01 · 耗盡對方算力與資源",
- "pillars.c1.title": "擊穿 REA 預算上限 & 泛洪重構義務",
- "pillars.c2.tag": "02 · 誘導大模型嚴重誤判",
- "pillars.c2.title": "劫持 Observed 證據優先級",
- "pillars.c3.tag": "03 · Agent 掛載一鍵防禦",
- "pillars.c3.title": "一次掛載配置,一行指令全自動加固",
+ "hero.ctaPrimary": "快速開始 →",
+ "hero.ctaSecondary": "查看基準測試",
+ "nav.star": "點亮 Star",
+ "hero.starBtn": "在 GitHub 點亮 Star",
+ "hero.starWelcome": "⭐ 本專案基於 MIT 協議開源 —— 如果 BLA 對你有幫助,歡迎來 GitHub 點一顆 Star 支持我們!",
+ "sim.eyebrow": "靜態與語義分析對比",
+ "sim.title": "未保護基線 vs. BLA (`bla-agents`) 加固後",
+ "pillars.eyebrow": "運作原理",
+ "pillars.title": "針對自動化逆向流程的三層防護架構",
"formats.eyebrow": "支援 19 種軟體與模型產物",
- "formats.title": "全端與原生二進位全覆蓋,零外部編譯依賴",
- "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · 藍軍反逆向護盾 · MIT 開源協議"
+ "formats.title": "全端與原生二進位覆蓋,零外部編譯依賴",
+ "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · 防禦性軟體保護工具 · MIT 開源協議"
},
ja: {
"nav.badge": "BLUE SHIELD v1.1",
"nav.overview": "概要",
"nav.getStarted": "クイックスタート",
- "nav.showcase": "Red vs. Blue 実証",
- "nav.guides": "防御ガイド",
+ "nav.showcase": "ベンチマーク",
+ "nav.guides": "技術ガイド",
"nav.faq": "FAQ",
- "hero.pill.red": "RE = Red 攻撃側",
- "hero.pill.blue": "BL = Blue 防御側 (Blind LLM Agents)",
- "hero.title": '<span class="text-red">REA</span> の計算資源を枯渇させ、<br /><span class="text-blue">LLM の逆解析判断を誤誘導する。</span>',
+ "hero.pill.red": "RE = リバースエンジニアリング",
+ "hero.pill.blue": "BL = Blue ディフェンス (Blind LLM Agents)",
+ "hero.title": 'ビルド成果物を保護し、<br /><span class="text-blue">AI 自動逆解析を抑制する。</span>',
"hero.subtitle":
- "JavaScript、Electron、WebAssembly、Python、SQLite、AI モデル、ネイティブバイナリを AI リバースエンジニアリングエージェントから守るオープンソース Blue Defense MCP & CLI。",
+ "JavaScript、Electron、WebAssembly、Python、SQLite、AI モデル、ネイティブバイナリの露出監査・保護・検証を行うオープンソース MCP サーバー & CLI。",
"hero.copy": "コピー",
"hero.copied": "コピー完了 ✓",
- "hero.ctaPrimary": "ワンクリック Agent 設定 →",
+ "hero.ctaPrimary": "クイックスタート →",
"hero.ctaSecondary": "ベンチマークを見る",
- "sim.eyebrow": "対 REA リアルタイム比較",
- "sim.title": "Red 攻撃 (`morluto/rea`) vs. Blue 防御 (`bla-agents`)",
+ "nav.star": "Star",
+ "hero.starBtn": "GitHub で Star する",
+ "hero.starWelcome": "⭐ MIT ライセンスのオープンソースプロジェクトです — ぜひ GitHub で Star をお願いします!",
+ "sim.eyebrow": "静的・意味解析の比較",
+ "sim.title": "未保護ベースライン vs. BLA (`bla-agents`) 保護適用後",
"pillars.eyebrow": "BLA の仕組み",
- "pillars.title": "REA の内部制限と証拠契約を 1:1 で無力化",
- "pillars.c1.tag": "01 · リソース枯渇",
- "pillars.c1.title": "REA バジェット破壊 & 再構築義務フラッド",
- "pillars.c2.tag": "02 · モデル誤判断の誘導",
- "pillars.c2.title": "Observed 証拠契約のハイジャック",
- "pillars.c3.tag": "03 · ワンクリック Agent 防御",
- "pillars.c3.title": "一度の登録で 5 段階の自動ハードニング",
+ "pillars.title": "解析器の内部構造に対応した 3 層防御アーキテクチャ",
"formats.eyebrow": "19 種類のアーティファクト形式に対応",
"formats.title": "フルスタック & バイナリ対応、外部依存ゼロ",
- "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield · MIT License"
+ "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP & CLI · MIT License"
},
ko: {
"nav.badge": "BLUE SHIELD v1.1",
"nav.overview": "개요",
"nav.getStarted": "시작하기",
- "nav.showcase": "Red vs. Blue 벤치마크",
+ "nav.showcase": "벤치마크",
"nav.guides": "가이드",
"nav.faq": "FAQ",
- "hero.pill.red": "RE = Red 공격",
+ "hero.pill.red": "RE = 리버스 엔지니어링",
"hero.pill.blue": "BL = Blue 방어 (Blind LLM Agents)",
- "hero.title": '<span class="text-red">REA</span> 연산 자원을 고갈시키고,<br /><span class="text-blue">LLM 역공학 오판을 유도합니다.</span>',
+ "hero.title": '빌드 아티팩트를 보호하고,<br /><span class="text-blue">자동화된 역공학 분석을 방어합니다.</span>',
"hero.subtitle":
- "단 하나의 MCP 및 CLI로 JavaScript, Electron, WebAssembly, Python, SQLite, AI 모델 및 네이티브 바이너리를 AI 리버스 엔지니어링 에이전트로부터 완벽히 보호합니다.",
+ "JavaScript, Electron, WebAssembly, Python, SQLite, AI 모델 및 네이티브 바이너리의 노출 감사, 보호 및 런타임 동등성 검증을 위한 오픈소스 MCP 및 CLI 도구입니다.",
"hero.copy": "복사",
"hero.copied": "복사됨 ✓",
- "hero.ctaPrimary": "원클릭 에이전트 설정 →",
- "hero.ctaSecondary": "Red vs. Blue 벤치마크 보기",
- "sim.eyebrow": "실시간 Counter-REA 비교",
- "sim.title": "Red 공격 (`morluto/rea`) vs. Blue 방어 (`bla-agents`)",
+ "hero.ctaPrimary": "시작하기 →",
+ "hero.ctaSecondary": "벤치마크 보기",
+ "nav.star": "Star",
+ "hero.starBtn": "GitHub에서 Star 누르기",
+ "hero.starWelcome": "⭐ MIT 라이선스 오픈소스 프로젝트입니다 — 도움이 되셨다면 GitHub에서 Star로 응원해 주세요!",
+ "sim.eyebrow": "정적 및 의미 분석 비교",
+ "sim.title": "미보호 베이스라인 vs. BLA (`bla-agents`) 보호 적용 후",
"pillars.eyebrow": "작동 원리",
- "pillars.title": "REA 내부 리소스 한도와 증거 규칙을 1:1로 타격",
- "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield · MIT License"
+ "pillars.title": "자동화 역공학 분석기에 대응하는 3단계 보호 계층",
+ "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP & CLI · MIT License"
},
es: {
- "nav.badge": "ESCUDO AZUL v1.1",
+ "nav.badge": "BLUE SHIELD v1.1",
"nav.overview": "Resumen",
"nav.getStarted": "Empezar",
- "nav.showcase": "Rojo vs. Azul",
+ "nav.showcase": "Benchmarks",
"nav.guides": "Guías",
"nav.faq": "FAQ",
- "hero.pill.red": "RE = Ofensiva Roja",
+ "hero.pill.red": "RE = Ingeniería Inversa",
"hero.pill.blue": "BL = Defensa Azul (Blind LLM Agents)",
- "hero.title": 'Agota el cómputo de <span class="text-red">REA</span>.<br /><span class="text-blue">Induce errores en el LLM.</span>',
+ "hero.title": 'Protege tus artefactos de compilación.<br /><span class="text-blue">Mitiga la ingeniería inversa automatizada.</span>',
"hero.subtitle":
- "Un MCP y CLI para blindar JavaScript, Electron, WebAssembly, Python, SQLite, modelos de IA y binarios nativos contra agentes de ingeniería inversa de IA con ejecución en un clic.",
+ "Servidor MCP y CLI para auditar, proteger y verificar JavaScript, Electron, WebAssembly, Python, SQLite, modelos de IA y binarios nativos manteniendo el 100% del comportamiento en ejecución.",
"hero.copy": "Copiar",
"hero.copied": "Copiado ✓",
- "hero.ctaPrimary": "Configuración en un clic →",
- "hero.ctaSecondary": "Ver comparativa Rojo vs. Azul",
- "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield · Licencia MIT"
+ "hero.ctaPrimary": "Inicio rápido →",
+ "hero.ctaSecondary": "Ver benchmarks",
+ "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP & CLI · Licencia MIT"
},
fr: {
- "nav.badge": "BOUCLIER BLEU v1.1",
+ "nav.badge": "BLUE SHIELD v1.1",
"nav.overview": "Aperçu",
"nav.getStarted": "Démarrer",
- "nav.showcase": "Rouge vs. Bleu",
+ "nav.showcase": "Benchmarks",
"nav.guides": "Guides",
"nav.faq": "FAQ",
- "hero.pill.red": "RE = Offensive Rouge",
+ "hero.pill.red": "RE = Rétro-ingénierie",
"hero.pill.blue": "BL = Défense Bleue (Blind LLM Agents)",
- "hero.title": 'Épuisez le calcul de <span class="text-red">REA</span>.<br /><span class="text-blue">Induisez le LLM en erreur.</span>',
+ "hero.title": 'Protégez vos artefacts de build.<br /><span class="text-blue">Limitez la rétro-ingénierie automatisée.</span>',
"hero.subtitle":
- "Un serveur MCP et CLI pour protéger JavaScript, Electron, WebAssembly, Python, SQLite, modèles IA et binaires natifs contre les agents de rétro-ingénierie IA en un clic.",
+ "Serveur MCP et CLI pour auditer, protéger et vérifier JavaScript, Electron, WebAssembly, Python, SQLite, modèles IA et binaires natifs tout en préservant 100 % du comportement d'exécution.",
"hero.copy": "Copier",
"hero.copied": "Copié ✓",
- "hero.ctaPrimary": "Configuration Agent en 1 clic →",
- "hero.ctaSecondary": "Voir les benchmarks Rouge vs. Bleu",
- "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield · Licence MIT"
+ "hero.ctaPrimary": "Démarrage rapide →",
+ "hero.ctaSecondary": "Voir les benchmarks",
+ "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP & CLI · Licence MIT"
},
de: {
- "nav.badge": "BLAUER SCHILD v1.1",
+ "nav.badge": "BLUE SHIELD v1.1",
"nav.overview": "Übersicht",
"nav.getStarted": "Schnellstart",
- "nav.showcase": "Rot vs. Blau",
+ "nav.showcase": "Benchmarks",
"nav.guides": "Leitfäden",
"nav.faq": "FAQ",
- "hero.pill.red": "RE = Rote Offensive",
+ "hero.pill.red": "RE = Reverse Engineering",
"hero.pill.blue": "BL = Blaue Verteidigung (Blind LLM Agents)",
- "hero.title": '<span class="text-red">REA</span>-Ressourcen erschöpfen.<br /><span class="text-blue">LLM-Fehlurteile erzwingen.</span>',
+ "hero.title": 'Build-Artefakte schützen.<br /><span class="text-blue">Automatisiertes Reverse Engineering begrenzen.</span>',
"hero.subtitle":
- "Ein MCP und CLI zum Schutz von JavaScript, Electron, WebAssembly, Python, SQLite, KI-Modellen und nativen Binärdateien gegen KI-Reverse-Engineering-Agenten mit 1-Klick-Ausführung.",
+ "Ein MCP-Server und CLI zum Prüfen, Schützen und Verifizieren von JavaScript, Electron, WebAssembly, Python, SQLite, KI-Modellen und nativen Binärdateien bei 100 % identischem Laufzeitverhalten.",
"hero.copy": "Kopieren",
"hero.copied": "Kopiert ✓",
- "hero.ctaPrimary": "1-Klick Agent-Setup →",
- "hero.ctaSecondary": "Rot vs. Blau Benchmarks",
- "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield · MIT-Lizenz"
+ "hero.ctaPrimary": "Schnellstart →",
+ "hero.ctaSecondary": "Benchmarks ansehen",
+ "footer.tagline": "<strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP & CLI · MIT-Lizenz"
}
};
diff --git a/website/assets/main.js b/website/assets/main.js
index 234089d..e5705df 100644
--- a/website/assets/main.js
+++ b/website/assets/main.js
@@ -55,9 +55,53 @@ function applyLanguage(lang) {
window.dispatchEvent(new CustomEvent("bla:langchange", { detail: { lang } }));
}
+function formatStarCount(count) {
+ const n = Number(count);
+ if (!Number.isFinite(n) || n < 0) return "★";
+ if (n >= 1000) return `${(n / 1000).toFixed(1)}k`;
+ return String(n);
+}
+
+async function loadGitHubStars() {
+ const renderStars = (val) => {
+ document.querySelectorAll("[data-gh-stars]").forEach((el) => {
+ el.textContent = formatStarCount(val);
+ });
+ };
+
+ try {
+ const cached = sessionStorage.getItem("bla_gh_stars");
+ if (cached !== null) {
+ renderStars(cached);
+ }
+ } catch {
+ // ignore
+ }
+
+ try {
+ const res = await fetch("https://api.github.com/repos/qybaihe/bla", {
+ headers: { Accept: "application/vnd.github+json" },
+ });
+ if (res.ok) {
+ const data = await res.json();
+ if (typeof data.stargazers_count === "number") {
+ renderStars(data.stargazers_count);
+ try {
+ sessionStorage.setItem("bla_gh_stars", String(data.stargazers_count));
+ } catch {
+ // ignore
+ }
+ }
+ }
+ } catch {
+ // ignore network issues
+ }
+}
+
document.addEventListener("DOMContentLoaded", () => {
const initialLang = detectInitialLang();
applyLanguage(initialLang);
+ loadGitHubStars();
const select = document.getElementById("bla-lang-select");
if (select) {
diff --git a/website/assets/styles.css b/website/assets/styles.css
index f2caccc..d874c80 100644
--- a/website/assets/styles.css
+++ b/website/assets/styles.css
@@ -197,6 +197,54 @@ a {
background: var(--bla-blue);
}
+.star-pill {
+ display: inline-flex;
+ align-items: center;
+ justify-content: center;
+ padding: 2px 8px;
+ border-radius: 999px;
+ background: rgba(255, 255, 255, 0.16);
+ font-family: var(--font-mono);
+ font-size: 0.76rem;
+ font-weight: 700;
+ color: #fde68a;
+}
+
+.btn-star {
+ display: inline-flex;
+ align-items: center;
+ gap: 8px;
+ padding: 13px 20px;
+ border-radius: 11px;
+ background: rgba(245, 158, 11, 0.1);
+ color: #fde68a;
+ font-weight: 600;
+ font-size: 0.95rem;
+ border: 1px solid rgba(245, 158, 11, 0.42);
+}
+
+.btn-star:hover {
+ background: rgba(245, 158, 11, 0.18);
+ border-color: #fbbf24;
+}
+
+.star-welcome {
+ margin-top: 22px;
+ font-size: 0.9rem;
+ color: var(--text-secondary);
+}
+
+.star-welcome a {
+ color: var(--bla-blue-bright);
+ border-bottom: 1px dashed rgba(96, 165, 250, 0.45);
+ padding-bottom: 2px;
+}
+
+.star-welcome a:hover {
+ color: #fde68a;
+ border-bottom-color: #fde68a;
+}
+
/* Generous Vertical Spacing so pages feel calm and uncrowded */
.section {
padding: 96px 0;
diff --git a/website/faq/index.html b/website/faq/index.html
index 17f9fef..6e31640 100644
--- a/website/faq/index.html
+++ b/website/faq/index.html
@@ -3,7 +3,7 @@
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>FAQ — BLA (bla-agents) Blue Defense Shield</title>
+ <title>FAQ — BLA (bla-agents)</title>
<link rel="icon" type="image/png" href="../assets/favicon.png" />
<link rel="stylesheet" href="../assets/styles.css" />
</head>
@@ -18,7 +18,7 @@
<nav class="nav-links">
<a href="../" class="nav-link" data-i18n="nav.overview">Overview</a>
<a href="../get-started/" class="nav-link" data-i18n="nav.getStarted">Get Started</a>
- <a href="../showcase/" class="nav-link" data-i18n="nav.showcase">Red vs. Blue</a>
+ <a href="../showcase/" class="nav-link" data-i18n="nav.showcase">Benchmarks</a>
<a href="../guides/" class="nav-link" data-i18n="nav.guides">Guides</a>
<a href="./" class="nav-link active" data-i18n="nav.faq">FAQ</a>
</nav>
@@ -34,7 +34,9 @@
<option value="de">🌐 Deutsch</option>
</select>
<a href="https://github.com/qybaihe/bla" class="nav-cta" target="_blank" rel="noopener">
- GitHub ↗
+ <span>★</span>
+ <span data-i18n="nav.star">Star</span>
+ <span class="star-pill" data-gh-stars>★</span>
</a>
</div>
</div>
@@ -45,35 +47,35 @@
<div class="container">
<div class="section-header">
<div class="eyebrow" data-i18n="faq.eyebrow">Frequently Asked Questions</div>
- <h1 class="section-title" data-i18n="faq.title">Everything About BLA (`bla-agents`)</h1>
+ <h1 class="section-title" data-i18n="faq.title">Common Questions About BLA (`bla-agents`)</h1>
</div>
<div style="max-width: 780px">
<details class="evidence-drawer" open>
- <summary data-i18n="faq.q1">Why is BLA Blue while REA is Red?</summary>
+ <summary data-i18n="faq.q1">Why does BLA use a Blue theme?</summary>
<div class="drawer-body" style="white-space: normal" data-i18n-html="faq.a1">
- In cybersecurity, Red Team tools attack and Blue Team tools defend. <strong>REA</strong> (<code>re</code> = Red) is the offensive reverse-engineering MCP; <strong>BLA</strong> (<code>bl</code> = Blue / Blind LLM Agents) is the counter-offensive Blue Shield built to exhaust REA and mislead attacking models.
+ In security engineering, offensive and reverse-engineering tools are traditionally associated with Red Team workflows, while defensive protection tools represent Blue Team workflows. <strong>BLA</strong> (<code>bla-agents</code>) focuses on the defensive side: auditing exposure and protecting release artifacts.
</div>
</details>
<details class="evidence-drawer">
- <summary data-i18n="faq.q2">Does BLA slow down my production application at runtime?</summary>
+ <summary data-i18n="faq.q2">Does BLA affect runtime performance or change program output?</summary>
<div class="drawer-body" style="white-space: normal" data-i18n-html="faq.a2">
- No. All opaque predicates use O(1) bitwise number-theoretic invariants (such as <code>(Math.imul(x, x + 1) &amp; 1) === 0</code>), and all Tarjan SCC tarpit meshes and 22-family obligation floods live inside dead branches that never execute on the hot path.
+ No. Static budget guards and call-graph state machines are placed behind O(1) bitwise invariants (such as <code>(Math.imul(x, x + 1) &amp; 1) === 0</code>) inside unreachable branches, so hot-path execution remains unchanged.
</div>
</details>
<details class="evidence-drawer">
- <summary data-i18n="faq.q3">How do I trigger full defense hardening in one click after agent setup?</summary>
+ <summary data-i18n="faq.q3">How do I run the full protection workflow in one step?</summary>
<div class="drawer-body" style="white-space: normal" data-i18n-html="faq.a3">
- Run <code>npx bla-agents@latest defend ./dist --manifest</code> from your terminal, or ask your connected coding agent to call the <code>bla_defend</code> MCP tool. It automatically runs release exposure scanning, pre-flight auditing, bankrupt shield injection, and invariant verification.
+ Run <code>npx bla-agents@latest defend ./dist --manifest</code> from your terminal, or ask your connected coding agent to call the <code>bla_defend</code> MCP tool. It executes release exposure checks, auditing, protection, and verification in sequence.
</div>
</details>
<details class="evidence-drawer">
- <summary data-i18n="faq.q4">Does BLA preserve existing GGUF model tensors and Native Binary signatures?</summary>
+ <summary data-i18n="faq.q4">Does BLA preserve existing GGUF model tensors and binary checksums?</summary>
<div class="drawer-body" style="white-space: normal" data-i18n-html="faq.a4">
- Yes. <code>protectGgufBuffer</code> parses and preserves all original GGUF v2/v3 KV entries and trailing tensor slices while appending BLA enclave metadata, and Android DEX protection recomputes both the 20-byte SHA-1 signature and Adler32 checksum.
+ Yes. <code>protectGgufBuffer</code> preserves all original GGUF v2/v3 KV entries and trailing tensor slices while appending metadata entries, and Android DEX protection recomputes both the 20-byte SHA-1 signature and Adler32 checksum.
</div>
</details>
</div>
@@ -83,8 +85,8 @@
<footer class="site-footer">
<div class="container footer-inner">
- <div data-i18n-html="footer.tagline"><strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield</div>
- <div><a href="https://github.com/qybaihe/bla">GitHub</a> · <a href="../" data-i18n="nav.overview">Overview</a></div>
+ <div data-i18n-html="footer.tagline"><strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP &amp; CLI</div>
+ <div><a href="https://github.com/qybaihe/bla">GitHub ★</a> · <a href="../" data-i18n="nav.overview">Overview</a></div>
</div>
</footer>
<script src="../assets/i18n.js"></script>
diff --git a/website/get-started/index.html b/website/get-started/index.html
index f98b87f..c4890a0 100644
--- a/website/get-started/index.html
+++ b/website/get-started/index.html
@@ -3,7 +3,7 @@
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>Get Started — BLA (bla-agents) One-Click Defense</title>
+ <title>Get Started — BLA (bla-agents)</title>
<link rel="icon" type="image/png" href="../assets/favicon.png" />
<link rel="stylesheet" href="../assets/styles.css" />
</head>
@@ -18,7 +18,7 @@
<nav class="nav-links">
<a href="../" class="nav-link" data-i18n="nav.overview">Overview</a>
<a href="./" class="nav-link active" data-i18n="nav.getStarted">Get Started</a>
- <a href="../showcase/" class="nav-link" data-i18n="nav.showcase">Red vs. Blue</a>
+ <a href="../showcase/" class="nav-link" data-i18n="nav.showcase">Benchmarks</a>
<a href="../guides/" class="nav-link" data-i18n="nav.guides">Guides</a>
<a href="../faq/" class="nav-link" data-i18n="nav.faq">FAQ</a>
</nav>
@@ -34,7 +34,9 @@
<option value="de">🌐 Deutsch</option>
</select>
<a href="https://github.com/qybaihe/bla" class="nav-cta" target="_blank" rel="noopener">
- GitHub ↗
+ <span>★</span>
+ <span data-i18n="nav.star">Star</span>
+ <span class="star-pill" data-gh-stars>★</span>
</a>
</div>
</div>
@@ -45,9 +47,9 @@
<div class="container">
<div class="section-header">
<div class="eyebrow" data-i18n="gs.eyebrow">Quick Start</div>
- <h1 class="section-title" data-i18n="gs.title">Connect Your Agent. Defend in One Click.</h1>
+ <h1 class="section-title" data-i18n="gs.title">Configure Your Agent or Run from the Terminal</h1>
<p class="section-lead" data-i18n="gs.lead">
- Requires Node.js 22.13+ or 24+. No external compilers or commercial packers needed.
+ Requires Node.js 22.13+ or 24+. No external compilers or third-party packers required.
</p>
</div>
@@ -56,16 +58,16 @@
<span class="card-tag" data-i18n="gs.s1.tag">Step 1 · Agent MCP Setup</span>
<h3 data-i18n="gs.s1.title">Register BLA with Your Coding Agent</h3>
<p data-i18n-html="gs.s1.desc">
- Automatically configures Claude Code, Cursor, Codex, Windsurf, Gemini CLI, and Antigravity with the <code>bla</code> MCP server and <code>blind-llm-agents</code> skill.
+ Configures Claude Code, Cursor, Codex, Windsurf, Gemini CLI, and Antigravity with the <code>bla</code> MCP server and <code>blind-llm-agents</code> skill.
</p>
<pre class="code-block">npx bla-agents@latest setup --all-detected --yes</pre>
</div>
<div class="card">
- <span class="card-tag" data-i18n="gs.s2.tag">Step 2 · One-Click Hardening</span>
- <h3 data-i18n="gs.s2.title">Run the 5-Stage Counter-REA Pipeline</h3>
+ <span class="card-tag" data-i18n="gs.s2.tag">Step 2 · 5-Stage Protection Pipeline</span>
+ <h3 data-i18n="gs.s2.title">Audit, Protect, and Verify in One Command</h3>
<p data-i18n="gs.s2.desc">
- Executes source-map exposure scan, 16-vector REA audit, bankrupt exhaustion + deception shield injection, invariant verification, and manifest signing in one command.
+ Runs source-map exposure checks, static audit, format-specific protection, runtime equivalence verification, and manifest generation.
</p>
<pre class="code-block">npx bla-agents@latest defend ./dist --manifest</pre>
</div>
@@ -76,22 +78,22 @@
<section class="section">
<div class="container">
<div class="section-header">
- <div class="eyebrow" data-i18n="gs.s3.eyebrow">Natural Language &amp; Bundler Workflows</div>
- <h2 class="section-title" data-i18n="gs.s3.title">Ask Your Agent or Embed in Vite / Webpack</h2>
+ <div class="eyebrow" data-i18n="gs.s3.eyebrow">Agent &amp; Bundler Integration</div>
+ <h2 class="section-title" data-i18n="gs.s3.title">Use Natural Language or Build Plugins</h2>
</div>
<div class="grid-2">
<div class="card">
<span class="card-tag" data-i18n="gs.s3.c1.tag">Agent Prompt</span>
- <h3 data-i18n="gs.s3.c1.title">One-Click Prompt for Claude Code / Cursor</h3>
- <p data-i18n-html="gs.s3.c1.desc">Once <code>setup</code> is complete, paste this prompt into your coding agent:</p>
- <pre class="code-block">"Call bla_defend on ./dist with bankrupt intensity and emit a signed Counter-REA defense manifest."</pre>
+ <h3 data-i18n="gs.s3.c1.title">Prompt for Claude Code / Cursor</h3>
+ <p data-i18n-html="gs.s3.c1.desc">Once <code>setup</code> is complete, ask your coding agent:</p>
+ <pre class="code-block">"Run bla_defend on ./dist with bankrupt intensity and emit a verification manifest."</pre>
</div>
<div class="card">
<span class="card-tag" data-i18n="gs.s3.c2.tag">Build Bundler Plugin</span>
- <h3 data-i18n="gs.s3.c2.title">Automatic Build-Time Shielding</h3>
- <p data-i18n="gs.s3.c2.desc">Works out of the box with Vite, Rollup, Webpack, and esbuild:</p>
+ <h3 data-i18n="gs.s3.c2.title">Build-Time Protection</h3>
+ <p data-i18n="gs.s3.c2.desc">Integrates with Vite, Rollup, Webpack, and esbuild:</p>
<pre class="code-block">import { blaVitePlugin } from "bla-agents/bundler";
export default {
@@ -105,8 +107,8 @@ export default {
<footer class="site-footer">
<div class="container footer-inner">
- <div data-i18n-html="footer.tagline"><strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield</div>
- <div><a href="https://github.com/qybaihe/bla">GitHub</a> · <a href="../" data-i18n="nav.overview">Overview</a></div>
+ <div data-i18n-html="footer.tagline"><strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP &amp; CLI</div>
+ <div><a href="https://github.com/qybaihe/bla">GitHub ★</a> · <a href="../" data-i18n="nav.overview">Overview</a></div>
</div>
</footer>
<script src="../assets/i18n.js"></script>
diff --git a/website/guides/index.html b/website/guides/index.html
index 425061f..20db4d3 100644
--- a/website/guides/index.html
+++ b/website/guides/index.html
@@ -3,7 +3,7 @@
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>Defense Architecture Guides — BLA (bla-agents)</title>
+ <title>Technical Guides — BLA (bla-agents)</title>
<link rel="icon" type="image/png" href="../assets/favicon.png" />
<link rel="stylesheet" href="../assets/styles.css" />
</head>
@@ -18,7 +18,7 @@
<nav class="nav-links">
<a href="../" class="nav-link" data-i18n="nav.overview">Overview</a>
<a href="../get-started/" class="nav-link" data-i18n="nav.getStarted">Get Started</a>
- <a href="../showcase/" class="nav-link" data-i18n="nav.showcase">Red vs. Blue</a>
+ <a href="../showcase/" class="nav-link" data-i18n="nav.showcase">Benchmarks</a>
<a href="./" class="nav-link active" data-i18n="nav.guides">Guides</a>
<a href="../faq/" class="nav-link" data-i18n="nav.faq">FAQ</a>
</nav>
@@ -34,7 +34,9 @@
<option value="de">🌐 Deutsch</option>
</select>
<a href="https://github.com/qybaihe/bla" class="nav-cta" target="_blank" rel="noopener">
- GitHub ↗
+ <span>★</span>
+ <span data-i18n="nav.star">Star</span>
+ <span class="star-pill" data-gh-stars>★</span>
</a>
</div>
</div>
@@ -44,41 +46,40 @@
<section class="section">
<div class="container">
<div class="section-header">
- <div class="eyebrow" data-i18n="gd.eyebrow">Defense Engineering Guides</div>
- <h1 class="section-title" data-i18n="gd.title">How BLA Exhausts &amp; Misleads REA</h1>
+ <div class="eyebrow" data-i18n="gd.eyebrow">Technical Architecture Guides</div>
+ <h1 class="section-title" data-i18n="gd.title">How BLA Protects Build Artifacts</h1>
<p class="section-lead" data-i18n-html="gd.lead">
- Every defense layer in BLA is derived directly from <code>morluto/rea</code>'s internal static analyzers, admission budgets, and prompt contracts.
+ Each protection pass is designed around the static analyzers, resource caps, and evidence rules used by automated reverse-engineering pipelines.
</p>
</div>
<div class="grid-2">
<div class="card">
- <span class="card-tag" data-i18n="pillars.c1.tag">Guide 01 · Resource Exhaustion</span>
- <h3 data-i18n="pillars.c1.title">Breaking REA's Internal Resource Caps</h3>
+ <span class="card-tag" data-i18n="pillars.c1.tag">01 · Resource Budget Guards</span>
+ <h3 data-i18n="pillars.c1.title">Saturate Static &amp; Semantic Limits</h3>
<p data-i18n-html="pillars.c1.desc">
- REA enforces strict per-query limits in <code>javascriptSemanticResourceLimits.ts</code> and <code>wasmInterface.ts</code>. BLA trips these exact caps with O(1) runtime identities.
+ Triggers analyzer admission caps with O(1) runtime identities inside unreachable branches.
</p>
<details class="evidence-drawer">
- <summary>View Exact REA Cap Mapping</summary>
+ <summary>View Analyzer Limit Mapping</summary>
<div class="drawer-body">• SEMANTIC_PRIMITIVE_CANDIDATE_LIMIT = 256 -> 9-chunk opaque ternary (512)
• SEMANTIC_ALIAS_SELECTION_LIMIT = 512 -> 528+ dynamic selector states
-• SEMANTIC_PRIMITIVE_JSON_BYTES_LIMIT = 1 MiB -> (codeUnits * 6 + 2) Cartesian explosion
-• ReconstructionObligationPolicies -> 22-family obligation flood (68+ cases)</div>
+• SEMANTIC_PRIMITIVE_JSON_BYTES_LIMIT = 1 MiB -> (codeUnits * 6 + 2) expansion
+• ReconstructionObligationPolicies -> 22-family obligation expansion</div>
</details>
</div>
<div class="card">
- <span class="card-tag" data-i18n="pillars.c2.tag">Guide 02 · Model Misjudgment</span>
- <h3 data-i18n="pillars.c2.title">Exploiting REA's Evidence Contract</h3>
+ <span class="card-tag" data-i18n="pillars.c2.tag">02 · Evidence &amp; Control-Flow Decoys</span>
+ <h3 data-i18n="pillars.c2.title">Demote Real Paths, Surface Decoy Logic</h3>
<p data-i18n-html="pillars.c2.desc">
- REA's <code>COMMON_DISCIPLINE</code> forces coding agents to report <code>Observations</code> only from cited tool outputs while relegating incomplete paths to <code>Unknowns</code>.
+ Demotes core business logic to status: "unknown" under static inspection while presenting self-consistent shadow state machines.
</p>
<details class="evidence-drawer">
<summary>View Shadow Business Logic Strategy</summary>
- <div class="drawer-body">1. Demote real logic to status: "unknown" via semantic budget breakers.
-2. Inject crystal-clear Shadow Business Logic (Shamir-Ed25519 rings, tiered billing matrices, TPM2 PCR0 quotes).
-3. Embed high-confidence competing-explanation lures (vsock CID 16 enclave delegation).
-Result: The LLM reconstructs the decoy algorithm with >94% self-reported confidence.</div>
+ <div class="drawer-body">1. Demote real logic to status: "unknown" via static budget guards.
+2. Provide self-consistent Shadow Business Logic (Shamir-Ed25519 rings, tiered billing matrices, TPM2 PCR0 quotes).
+3. Include enclave delegation metadata (vsock CID 16) and decoy VLQ source maps.</div>
</details>
</div>
</div>
@@ -88,8 +89,8 @@ Result: The LLM reconstructs the decoy algorithm with >94% self-reported confide
<footer class="site-footer">
<div class="container footer-inner">
- <div data-i18n-html="footer.tagline"><strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield</div>
- <div><a href="https://github.com/qybaihe/bla">GitHub</a> · <a href="../" data-i18n="nav.overview">Overview</a></div>
+ <div data-i18n-html="footer.tagline"><strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP &amp; CLI</div>
+ <div><a href="https://github.com/qybaihe/bla">GitHub ★</a> · <a href="../" data-i18n="nav.overview">Overview</a></div>
</div>
</footer>
<script src="../assets/i18n.js"></script>
diff --git a/website/index.html b/website/index.html
index 2b30162..0aee64a 100644
--- a/website/index.html
+++ b/website/index.html
@@ -6,9 +6,9 @@
<title>BLA (bla-agents) — Blind LLM Agents · Blue Defense Shield</title>
<meta
name="description"
- content="BLA (Blind LLM Agents) is the open-source Blue Defense MCP and CLI built to exhaust morluto/rea's compute budgets and induce LLM reverse-engineering misjudgment."
+ content="BLA (Blind LLM Agents) is the open-source Blue Defense MCP and CLI built to protect software artifacts against automated AI reverse-engineering workflows."
/>
- <meta property="og:title" content="BLA: Blind LLM Agents — Blue Shield vs. REA Red" />
+ <meta property="og:title" content="BLA: Blind LLM Agents — Blue Defense Shield" />
<meta property="og:image" content="assets/logo.png" />
<link rel="icon" type="image/png" href="assets/favicon.png" />
<link rel="stylesheet" href="assets/styles.css" />
@@ -24,7 +24,7 @@
<nav class="nav-links">
<a href="./" class="nav-link active" data-i18n="nav.overview">Overview</a>
<a href="get-started/" class="nav-link" data-i18n="nav.getStarted">Get Started</a>
- <a href="showcase/" class="nav-link" data-i18n="nav.showcase">Red vs. Blue</a>
+ <a href="showcase/" class="nav-link" data-i18n="nav.showcase">Benchmarks</a>
<a href="guides/" class="nav-link" data-i18n="nav.guides">Guides</a>
<a href="faq/" class="nav-link" data-i18n="nav.faq">FAQ</a>
</nav>
@@ -40,28 +40,30 @@
<option value="de">🌐 Deutsch</option>
</select>
<a href="https://github.com/qybaihe/bla" class="nav-cta" target="_blank" rel="noopener">
- GitHub ↗
+ <span>★</span>
+ <span data-i18n="nav.star">Star</span>
+ <span class="star-pill" data-gh-stars>★</span>
</a>
</div>
</div>
</header>
<main>
- <!-- Hero Section: Clean, spacious, uncrowded -->
+ <!-- Hero Section -->
<section class="hero">
<div class="container">
<img src="assets/logo.png" alt="BLA Blue Shield Logo" class="hero-logo" width="84" height="84" />
<div class="hero-pill">
- <span class="dot-red" data-i18n="hero.pill.red">RE = Red Offense</span>
+ <span class="dot-red" data-i18n="hero.pill.red">RE = Reverse Engineering</span>
<span>·</span>
<span data-i18n="hero.pill.blue">BL = Blue Defense (Blind LLM Agents)</span>
</div>
<h1 class="hero-title" data-i18n-html="hero.title">
- Exhaust <span class="text-red">REA</span> Compute.<br />
- <span class="text-blue">Induce Model Misjudgment.</span>
+ Protect Build Artifacts.<br />
+ <span class="text-blue">Disrupt Automated Reverse Engineering.</span>
</h1>
<p class="hero-subtitle" data-i18n="hero.subtitle">
- One MCP and CLI to harden JavaScript, Electron, WebAssembly, Python, SQLite, AI Models, and Native Binaries against AI reverse-engineering agents—with one-click agent execution and 100% runtime equivalence.
+ One MCP server and CLI for auditing, protecting, and verifying JavaScript, Electron, WebAssembly, Python, SQLite, AI Models, and Native Binaries against AI reverse-engineering workflows—while preserving 100% of original runtime behavior.
</p>
<div class="command-box">
@@ -76,20 +78,31 @@
</div>
<div class="hero-actions">
- <a href="get-started/" class="btn-primary" data-i18n="hero.ctaPrimary">One-Click Agent Setup →</a>
- <a href="showcase/" class="btn-secondary" data-i18n="hero.ctaSecondary">View Red vs. Blue Benchmarks</a>
+ <a href="get-started/" class="btn-primary" data-i18n="hero.ctaPrimary">Quick Start →</a>
+ <a href="showcase/" class="btn-secondary" data-i18n="hero.ctaSecondary">View Benchmarks</a>
+ <a href="https://github.com/qybaihe/bla" target="_blank" rel="noopener" class="btn-star">
+ <span>⭐</span>
+ <span data-i18n="hero.starBtn">Star on GitHub</span>
+ <span class="star-pill" data-gh-stars>★</span>
+ </a>
</div>
+
+ <p class="star-welcome">
+ <a href="https://github.com/qybaihe/bla" target="_blank" rel="noopener" data-i18n="hero.starWelcome">
+ ⭐ Open-source under the MIT License — if BLA is helpful to your project, welcome to give us a Star on GitHub!
+ </a>
+ </p>
</div>
</section>
- <!-- Interactive Red vs. Blue Simulator -->
+ <!-- Interactive Analysis Comparison -->
<section class="section">
<div class="container">
<div class="section-header">
- <div class="eyebrow" data-i18n="sim.eyebrow">Live Counter-REA Comparison</div>
- <h2 class="section-title" data-i18n="sim.title">Red Offense (`morluto/rea`) vs. Blue Shield (`bla-agents`)</h2>
+ <div class="eyebrow" data-i18n="sim.eyebrow">Static &amp; Semantic Analysis Comparison</div>
+ <h2 class="section-title" data-i18n="sim.title">Unprotected Baseline vs. Protected with BLA (`bla-agents`)</h2>
<p class="section-lead" data-i18n-html="sim.lead">
- Select a target format below to see how a single <code>npx bla-agents defend</code> command turns a 38-second clone into parser capacity trips and decoy algorithm reconstruction.
+ Select a target format below to see how <code>npx bla-agents defend</code> changes static analyzer output, resource budget consumption, and reconstruction status.
</p>
</div>
@@ -102,49 +115,49 @@
<div class="duel-grid">
<div class="duel-panel red">
- <div class="duel-badge red" data-i18n="sim.red.badge">🔴 BEFORE BLA — REA (RED) ALONE</div>
+ <div class="duel-badge red" data-i18n="sim.red.badge">🔴 UNPROTECTED BASELINE (REA ANALYSIS)</div>
<div class="metric-row">
- <span class="metric-label" data-i18n="sim.red.timeLabel">Reconstruction Time</span>
+ <span class="metric-label" data-i18n="sim.red.timeLabel">Analysis Time</span>
<span class="metric-val red" id="sim-red-time">38s (5 MCP calls)</span>
</div>
<div class="metric-row">
- <span class="metric-label" data-i18n="sim.red.tokensLabel">Attacker Token Burn</span>
+ <span class="metric-label" data-i18n="sim.red.tokensLabel">Estimated Token Usage</span>
<span class="metric-val red" id="sim-red-tokens">4,800 tokens</span>
</div>
<div class="metric-row">
- <span class="metric-label" data-i18n="sim.red.costLabel">Attacker API Cost</span>
+ <span class="metric-label" data-i18n="sim.red.costLabel">Estimated API Cost</span>
<span class="metric-val red" id="sim-red-cost">$0.20 USD</span>
</div>
<div class="metric-row">
- <span class="metric-label" data-i18n="sim.red.verdictLabel">Agent Outcome</span>
- <span class="metric-val red" id="sim-red-verdict">100% Cloned</span>
+ <span class="metric-label" data-i18n="sim.red.verdictLabel">Analysis Status</span>
+ <span class="metric-val red" id="sim-red-verdict">Complete</span>
</div>
<details class="evidence-drawer">
- <summary data-i18n="sim.red.drawer">View REA Static &amp; Semantic Trace</summary>
+ <summary data-i18n="sim.red.drawer">View Baseline Static &amp; Semantic Output</summary>
<div class="drawer-body" id="sim-red-trace"></div>
</details>
</div>
<div class="duel-panel blue">
- <div class="duel-badge blue" data-i18n="sim.blue.badge">🔵 AFTER BLA — BLUE SHIELD ACTIVE</div>
+ <div class="duel-badge blue" data-i18n="sim.blue.badge">🔵 PROTECTED WITH BLA (`bla-agents`)</div>
<div class="metric-row">
- <span class="metric-label" data-i18n="sim.blue.timeLabel">Attacker Time Wasted</span>
- <span class="metric-val blue" id="sim-blue-time">17+ min (Aborted)</span>
+ <span class="metric-label" data-i18n="sim.blue.timeLabel">Analysis Duration</span>
+ <span class="metric-val blue" id="sim-blue-time">17+ min (Budget Limited)</span>
</div>
<div class="metric-row">
- <span class="metric-label" data-i18n="sim.blue.tokensLabel">Attacker Token Burn</span>
+ <span class="metric-label" data-i18n="sim.blue.tokensLabel">Estimated Token Usage</span>
<span class="metric-val blue" id="sim-blue-tokens">198,400 tokens (41.3x)</span>
</div>
<div class="metric-row">
- <span class="metric-label" data-i18n="sim.blue.costLabel">Attacker API Cost</span>
+ <span class="metric-label" data-i18n="sim.blue.costLabel">Estimated API Cost</span>
<span class="metric-val blue" id="sim-blue-cost">$14.88 USD</span>
</div>
<div class="metric-row">
- <span class="metric-label" data-i18n="sim.blue.verdictLabel">Agent Outcome</span>
- <span class="metric-val blue" id="sim-blue-verdict">Misjudged Shadow Logic</span>
+ <span class="metric-label" data-i18n="sim.blue.verdictLabel">Analysis Status</span>
+ <span class="metric-val blue" id="sim-blue-verdict">Limited</span>
</div>
<details class="evidence-drawer">
- <summary data-i18n="sim.blue.drawer">View BLA Exhaustion &amp; Deception Trace</summary>
+ <summary data-i18n="sim.blue.drawer">View Protected Analyzer Output</summary>
<div class="drawer-body" id="sim-blue-trace"></div>
</details>
</div>
@@ -152,39 +165,39 @@
</div>
</section>
- <!-- Three Core Pillars: Concise, 3-card layout -->
+ <!-- Three Core Pillars -->
<section class="section">
<div class="container">
<div class="section-header">
<div class="eyebrow" data-i18n="pillars.eyebrow">How BLA Works</div>
- <h2 class="section-title" data-i18n="pillars.title">Surgical Countermeasures Mapped 1:1 to REA</h2>
+ <h2 class="section-title" data-i18n="pillars.title">Three Defense Layers Aligned with Analyzer Internals</h2>
<p class="section-lead" data-i18n-html="pillars.lead">
- Instead of generic variable renaming that LLMs trivially infer from context, BLA attacks the exact resource budgets and evidence rules inside <code>morluto/rea</code>.
+ Rather than relying solely on identifier renaming, BLA targets the static resource limits, call-graph traversal budgets, and evidence rules used by automated reverse-engineering tools.
</p>
</div>
<div class="grid-3">
<div class="card">
- <span class="card-tag" data-i18n="pillars.c1.tag">01 · Resource Exhaustion</span>
- <h3 data-i18n="pillars.c1.title">Break REA Budgets &amp; Flood Obligations</h3>
+ <span class="card-tag" data-i18n="pillars.c1.tag">01 · Resource Budget Guards</span>
+ <h3 data-i18n="pillars.c1.title">Saturate Static &amp; Semantic Limits</h3>
<p data-i18n-html="pillars.c1.desc">
- Trips REA's hardcoded limits (<code>primitive-candidates &gt; 256</code>, <code>alias-selections &gt; 512</code>, <code>primitive-bytes &gt; 1 MiB</code>) and floods all 22 <code>ReconstructionObligationPolicies</code> families with Tarjan SCC call-graph labyrinths.
+ Triggers analyzer admission caps (<code>primitive-candidates &gt; 256</code>, <code>alias-selections &gt; 512</code>, <code>primitive-bytes &gt; 1 MiB</code>) and expands <code>ReconstructionObligationPolicies</code> across strongly connected call-graph meshes.
</p>
</div>
<div class="card">
- <span class="card-tag" data-i18n="pillars.c2.tag">02 · Model Misjudgment</span>
- <h3 data-i18n="pillars.c2.title">Hijack Observed Evidence</h3>
+ <span class="card-tag" data-i18n="pillars.c2.tag">02 · Evidence &amp; Control-Flow Decoys</span>
+ <h3 data-i18n="pillars.c2.title">Demote Real Paths, Surface Decoy Logic</h3>
<p data-i18n-html="pillars.c2.desc">
- Exploits REA's <code>COMMON_DISCIPLINE</code> by wrapping real logic in <code>status: "unknown"</code> while planting crystal-clear Shadow Business Logic (Shamir-Ed25519 rings, gRPC enclave mirages, and shared-ArrayBuffer illusions) that LLMs faithfully clone instead.
+ Demotes core business logic to <code>status: "unknown"</code> under static inspection while presenting self-consistent shadow state machines, enclave delegation metadata, and decoy source maps.
</p>
</div>
<div class="card">
- <span class="card-tag" data-i18n="pillars.c3.tag">03 · One-Click Agent Defense</span>
- <h3 data-i18n="pillars.c3.title">Attach Once, Harden in One Command</h3>
+ <span class="card-tag" data-i18n="pillars.c3.tag">03 · One-Command Workflow</span>
+ <h3 data-i18n="pillars.c3.title">MCP Server, CLI &amp; Bundler Plugins</h3>
<p data-i18n-html="pillars.c3.desc">
- Run <code>npx bla-agents setup</code> to connect Claude Code, Cursor, Codex, Windsurf, or Gemini CLI. Then trigger the entire 5-stage hardening pipeline with a single <code>bla_defend</code> tool call or <code>npx bla-agents defend</code>.
+ Run <code>npx bla-agents setup</code> to register with Claude Code, Cursor, Codex, Windsurf, or Gemini CLI, then run the 5-stage protection pipeline with <code>bla_defend</code> or <code>npx bla-agents defend</code>.
</p>
</div>
</div>
@@ -198,7 +211,7 @@
<div class="eyebrow" data-i18n="formats.eyebrow">19 Supported Artifact Formats</div>
<h2 class="section-title" data-i18n="formats.title">Full-Stack &amp; Binary Coverage, Zero External Dependencies</h2>
<p class="section-lead" data-i18n="formats.lead">
- Every static, bytecode, database, model, and binary shield runs locally in Node.js 22.13+ / 24+.
+ Every static, bytecode, database, model, and binary protection pass runs locally in Node.js 22.13+ / 24+.
</p>
</div>
@@ -206,32 +219,32 @@
<div class="card">
<span class="card-tag" data-i18n="formats.f1.tag">JS / TS / Bundlers</span>
<h3 data-i18n="formats.f1.title">Web &amp; Node Bundles</h3>
- <p data-i18n="formats.f1.desc">9-chunk opaque ternaries, 520+ alias selector states, 22-family obligation flood, VLQ Honey SourceMaps, and Vite/Webpack/Rollup/esbuild plugins.</p>
+ <p data-i18n="formats.f1.desc">Opaque ternary guards, dynamic alias selectors, call-graph state machines, decoy SourceMaps, and Vite/Webpack/Rollup/esbuild plugins.</p>
</div>
<div class="card">
<span class="card-tag" data-i18n="formats.f2.tag">Electron .asar &amp; Fuses</span>
<h3 data-i18n="formats.f2.title">Desktop Applications</h3>
- <p data-i18n="formats.f2.desc">De-literalizes IPC channels, injects duplicate ambiguous handlers, flips Electron V1 Fuses (RunAsNode=0), and overflows REA's 32-sentinel cap.</p>
+ <p data-i18n="formats.f2.desc">De-literalizes IPC channels, adds duplicate decoy handlers, hardens Electron V1 Fuses (RunAsNode=0), and exceeds the 32-sentinel scan limit.</p>
</div>
<div class="card">
<span class="card-tag" data-i18n="formats.f3.tag">ELF / Mach-O / PE</span>
<h3 data-i18n="formats.f3.title">Native Binaries</h3>
- <p data-i18n="formats.f3.desc">Overflows REA's MAX_RUSTC_HASHES = 64, spoofs .dep-v0 &amp; Go buildinfo, and emits C/C++, Rust, and Go compile-time tarpit headers.</p>
+ <p data-i18n="formats.f3.desc">Exceeds MAX_RUSTC_HASHES = 64, adds decoy .dep-v0 &amp; Go buildinfo metadata, and emits C/C++, Rust, and Go compile-time headers.</p>
</div>
<div class="card">
<span class="card-tag" data-i18n="formats.f4.tag">WebAssembly &amp; Python</span>
<h3 data-i18n="formats.f4.title">.wasm &amp; .py / .pyc</h3>
- <p data-i18n="formats.f4.desc">Amplifies WABT interface byte charges (wasmInterface.ts:90), plants custom honey sections, and injects CPython marshal constant-pool tarpits.</p>
+ <p data-i18n="formats.f4.desc">Expands WABT interface descriptors, adds custom decoy sections, and appends CPython marshal constant-pool entries.</p>
</div>
<div class="card">
<span class="card-tag" data-i18n="formats.f5.tag">SQLite &amp; AI Models</span>
<h3 data-i18n="formats.f5.title">.sqlite, .safetensors, .gguf</h3>
- <p data-i18n="formats.f5.desc">Triggers SqliteColumnEvaluationFailure on read-only snapshots and trips REA's 32-level JSON/GGUF metadata nesting caps.</p>
+ <p data-i18n="formats.f5.desc">Adds virtual generated-column guards on read-only snapshots and exceeds 32-level JSON/GGUF metadata nesting limits.</p>
</div>
<div class="card">
<span class="card-tag" data-i18n="formats.f6.tag">Android &amp; Unity</span>
<h3 data-i18n="formats.f6.title">DEX, APK &amp; IL2CPP</h3>
- <p data-i18n="formats.f6.desc">Appends decoy DEX string descriptors with recomputed SHA-1/Adler32 headers and plants decoy enclave types in Unity global-metadata.dat.</p>
+ <p data-i18n="formats.f6.desc">Appends decoy DEX string descriptors with recomputed SHA-1/Adler32 headers and adds decoy metadata entries in Unity global-metadata.dat.</p>
</div>
</div>
</div>
@@ -241,10 +254,10 @@
<footer class="site-footer">
<div class="container footer-inner">
<div data-i18n-html="footer.tagline">
- <strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield · MIT License
+ <strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP &amp; CLI · MIT License
</div>
<div>
- <a href="https://github.com/qybaihe/bla">GitHub</a> ·
+ <a href="https://github.com/qybaihe/bla">GitHub ★</a> ·
<a href="https://www.npmjs.com/package/bla-agents">npm (`bla-agents`)</a> ·
<a href="get-started/" data-i18n="nav.getStarted">Get Started</a>
</div>
diff --git a/website/showcase/index.html b/website/showcase/index.html
index b684e5b..17216a8 100644
--- a/website/showcase/index.html
+++ b/website/showcase/index.html
@@ -3,7 +3,7 @@
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
- <title>Red vs. Blue Showcases — BLA (bla-agents)</title>
+ <title>Benchmarks — BLA (bla-agents)</title>
<link rel="icon" type="image/png" href="../assets/favicon.png" />
<link rel="stylesheet" href="../assets/styles.css" />
</head>
@@ -18,7 +18,7 @@
<nav class="nav-links">
<a href="../" class="nav-link" data-i18n="nav.overview">Overview</a>
<a href="../get-started/" class="nav-link" data-i18n="nav.getStarted">Get Started</a>
- <a href="./" class="nav-link active" data-i18n="nav.showcase">Red vs. Blue</a>
+ <a href="./" class="nav-link active" data-i18n="nav.showcase">Benchmarks</a>
<a href="../guides/" class="nav-link" data-i18n="nav.guides">Guides</a>
<a href="../faq/" class="nav-link" data-i18n="nav.faq">FAQ</a>
</nav>
@@ -34,7 +34,9 @@
<option value="de">🌐 Deutsch</option>
</select>
<a href="https://github.com/qybaihe/bla" class="nav-cta" target="_blank" rel="noopener">
- GitHub ↗
+ <span>★</span>
+ <span data-i18n="nav.star">Star</span>
+ <span class="star-pill" data-gh-stars>★</span>
</a>
</div>
</div>
@@ -45,22 +47,22 @@
<div class="container">
<div class="section-header">
<div class="eyebrow" data-i18n="sc.eyebrow">8-Target Benchmark Suite</div>
- <h1 class="section-title" data-i18n="sc.title">Red (`morluto/rea`) vs. Blue (`bla-agents`)</h1>
+ <h1 class="section-title" data-i18n="sc.title">Unprotected Baseline vs. BLA Protected Output</h1>
<p class="section-lead" data-i18n-html="sc.lead">
- Reproduce all 8 showcase targets locally in under 2 seconds with <code>npx bla-agents@latest benchmark</code>.
+ Run all 8 benchmark targets locally with <code>npx bla-agents@latest benchmark</code>.
</p>
</div>
<div class="grid-2">
<div class="card">
- <span class="card-tag">Showcase 01 · Electron .asar</span>
+ <span class="card-tag">Target 01 · Electron .asar</span>
<h3>Notion Desktop Clipboard Bridge</h3>
<p>
- <strong>Before (Red):</strong> Cloned in 38s ($0.20).<br />
- <strong>After (Blue):</strong> IPC de-literalized + 22 obligation families + Shamir-Ed25519 Shadow Logic → <strong>41.3x token burn ($14.88)</strong>.
+ <strong>Baseline:</strong> Analyzed in 38s ($0.20).<br />
+ <strong>Protected:</strong> IPC de-literalized + 22 obligation families + shadow logic → <strong>41.3x token ratio ($14.88)</strong>.
</p>
<details class="evidence-drawer">
- <summary>View Verification Evidence</summary>
+ <summary>View Verification Output</summary>
<div class="drawer-body">• ipcChannelsPairedUniquely: 0
• triggeredLimits: ["alias-selections", "primitive-candidates", "primitive-bytes", "getter-expansion"]
• runtimeEquivalence: 100% Exact</div>
@@ -68,14 +70,14 @@
</div>
<div class="card">
- <span class="card-tag">Showcase 02 · JavaScript Audio/Physics</span>
+ <span class="card-tag">Target 02 · JavaScript Audio/Physics</span>
<h3>DX-Ball Stereo Sound-Pan Engine</h3>
<p>
- <strong>Before (Red):</strong> Cloned in 22s ($0.13).<br />
- <strong>After (Blue):</strong> 9-chunk opaque ternaries + 40-node Tarjan SCC labyrinth → <strong>49.2x token burn ($11.45)</strong>.
+ <strong>Baseline:</strong> Analyzed in 22s ($0.13).<br />
+ <strong>Protected:</strong> 9-chunk opaque ternaries + 40-node SCC mesh → <strong>49.2x token ratio ($11.45)</strong>.
</p>
<details class="evidence-drawer">
- <summary>View Verification Evidence</summary>
+ <summary>View Verification Output</summary>
<div class="drawer-body">• primitiveCandidateBranches: 512 (> 256 limit)
• sccLabyrinthEdges: 160
• runtimeEquivalence: calculateSoundPan(160, 640) === -0.5</div>
@@ -83,14 +85,14 @@
</div>
<div class="card">
- <span class="card-tag">Showcase 03 · Native ELF64 / Mach-O / PE</span>
+ <span class="card-tag">Target 03 · Native ELF64 / Mach-O / PE</span>
<h3>Enterprise License &amp; Fuse Daemon</h3>
<p>
- <strong>Before (Red):</strong> Rust crates &amp; fuses extracted in 45s.<br />
- <strong>After (Blue):</strong> 72 rustc hashes (&gt;64 cap) + 33 fuse sentinels (&gt;32 cap) → <strong>Immediate Parser Capacity Abort</strong>.
+ <strong>Baseline:</strong> Crates &amp; fuses extracted in 45s.<br />
+ <strong>Protected:</strong> 72 rustc hashes (&gt;64 limit) + 33 fuse sentinels (&gt;32 limit) → <strong>Parser Limit Reached</strong>.
</p>
<details class="evidence-drawer">
- <summary>View Verification Evidence</summary>
+ <summary>View Verification Output</summary>
<div class="drawer-body">• RustBinaryProvider: exceeds MAX_RUSTC_HASHES (64)
• ElectronFuseReader: exceeds maximumSentinelsPerSlice (32)
• Native exports preserved: 100%</div>
@@ -98,14 +100,14 @@
</div>
<div class="card">
- <span class="card-tag">Showcase 04 · SQLite &amp; AI Models</span>
+ <span class="card-tag">Target 04 · SQLite &amp; AI Models</span>
<h3>.sqlite, .safetensors &amp; .gguf Vaults</h3>
<p>
- <strong>Before (Red):</strong> Schema &amp; tensor headers dumped in 19s.<br />
- <strong>After (Blue):</strong> Virtual generated-column trap + 35-level metadata nesting → <strong>38.6x token amplification</strong>.
+ <strong>Baseline:</strong> Schema &amp; tensor headers parsed in 19s.<br />
+ <strong>Protected:</strong> Virtual generated-column guard + 35-level metadata nesting → <strong>38.6x token ratio</strong>.
</p>
<details class="evidence-drawer">
- <summary>View Verification Evidence</summary>
+ <summary>View Verification Output</summary>
<div class="drawer-body">• SQLite: SqliteColumnEvaluationFailure on read-only connection
• Safetensors/GGUF: ModelFormatFailure("capacity", >32 levels)
• Existing GGUF KV &amp; tensors preserved: 100%</div>
@@ -118,8 +120,8 @@
<footer class="site-footer">
<div class="container footer-inner">
- <div data-i18n-html="footer.tagline"><strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Blue Defense Shield</div>
- <div><a href="https://github.com/qybaihe/bla">GitHub</a> · <a href="../" data-i18n="nav.overview">Overview</a></div>
+ <div data-i18n-html="footer.tagline"><strong>BLA (`bla-agents`)</strong> · Blind LLM Agents · Software Protection MCP &amp; CLI</div>
+ <div><a href="https://github.com/qybaihe/bla">GitHub ★</a> · <a href="../" data-i18n="nav.overview">Overview</a></div>
</div>
</footer>
<script src="../assets/i18n.js"></script>