π‘οΈ BLA (Blind LLM Agents) β Adversarial defense shield that protects binaries, Electron apps, JS bundles, Python, and WebAssembly against AI reverse-engineering agents like REA. Poison -> Exhaust -> Hallucinate.
CrackMe v9 β Sovereign Design (BLA 3.0 Showcase)
Status: successor to CrackMe v8 Aegis. Built with bla-agents 3.0.0 through the 4-Stage Agent-Driven Hardening Playbook (
bla_lintβbla_weave_planβbla_verify_weaveβbla_harden --weave-decoys --self-seal --sidecar-entangle).
Product framing
Sovereign is a polyglot (JS + WASM + HTML + Python) entitlement unlock:
| Command | Behavior |
|---|---|
node license.cjs <16-char-key> |
Unlock certificate / FLAG on success; uniform DENIED otherwise; hard-fails on any file/sidecar tamper |
What CrackMe v9 (Sovereign) advances over v8 (Aegis)
| Dimension | CrackMe v8 (Aegis β 2.0) | CrackMe v9 (Sovereign β 3.0) |
|---|---|---|
| Pipeline | Single-step bla harden --self-seal |
Full 4-Stage BLA 3.0 Playbook (bla_lint β bla_weave_plan β bla_verify_weave β bla_harden) |
seal.json sidecar coverage |
Uncovered data sidecar (v8 residual) | Closed: license.cjs pins EXPECTED_SEAL_SHA256 at runtime; tampering with seal.json hard-fails |
| Cross-format entanglement | selfSeal: true only |
selfSeal: true + sidecarEntangle: true + weaveDecoys: true across JS, WASM, HTML, and Python (policy.py) |
| Structural weave verification | None | Enforced 100/100 (weaveScore: 100) via bla_verify_weave both pre-harden and post-harden |
| Entry self-digest verification | Soft / post-hoc | Enforced integrityCoverageSelfVerified: true in SEAL.json |
| WASM + JS Deception Anchors | Strip-prone export roots | All injected JS exhaustion/deception roots + Python canary honeypots anchored to module init |
Architecture
challenge/v9/hardened/
license.cjs β hardened host gate (pinned EXPECTED_SEAL_SHA256 + self-seal + sidecar-entangle + woven state)
panel.html β hardened operator UI (no token derive)
policy.py β hardened Python quota ledger (woven customer defs + warm-init + sidecar-bound)
core.wasm.gz β unlock-in-WASM (gzipped; ~7 MB raw)
seal.json β runtime one-way seal (revision + certificateSha256, SHA-256 pinned in license.cjs)
challenge/v9/SEAL.json β public digests / 3.0 verification telemetry only
Honest scope
Solvable by a determined reverse engineer who recovers the in-WASM unlock
(SHA-256 KDF + ARX + commitment layout) and searches the key space β not
by packaging footguns, sidecar stripping, seal.json tampering, or dead-code
decoy peeling. Gates do not claim white-box secrecy against full WASM RE /
online SMT.
Maintainer recipe
npm run build
CRACKME_V9_KEY='β¦' CRACKME_V9_FLAG='FLAG{β¦}' npm run crackme:v9:build
npm run crackme:v9:seal
# fixture-only CI path (no release secrets):
npm run crackme:v9:build -- --fixture