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.

953 B
Raw Blame

Contributing to BLA (bla-shield)

Thank you for helping software developers fight back against automated AI reverse engineering!

Development Workflow

# Install dependencies
npm install

# Build TypeScript to dist/
npm run build

# Run the Vitest test suite
npm test

# Run the empirical Counter-REA benchmark
node dist/cli.js benchmark

Architecture Guidelines

  1. Zero Runtime Side Effects: Every AST transformation, opaque predicate, and binary section extension must be 100% side-effect free at runtime. Every shield must be verified by executing both the original and protected artifacts and asserting identical return values.
  2. Deterministic Builds: Always thread SeededPrng (--seed) through new shield passes so builds are reproducible in CI/CD pipelines.
  3. Target Real Analyzer Limits: When adding a new countermeasure, document the exact static/semantic analyzer heuristic or resource limit it defeats.