BLA: A Free Tool That Plants Decoys to Slow AI Agents That Reverse Engineer Your Software

BLA is a free tool that fills a finished program with decoys so AI agents struggle to take it apart. In its author’s tests, an AI reverse-engineering toolkit called REA hit an error or a limit on 10 of 13 kinds of test file.
cyber-security
artificial-intelligence
Author

Kabui, Charles

Published

2026-11-03

Keywords

reverse-engineering, software-protection, obfuscation, ai-agents, adversarial-defense

Download as PDF

BLA, short for Blind LLM Agents, is a free tool that software makers run on a finished program so AI agents that reverse-engineer it waste time and money. Reverse engineering means working out how software works without its source code, and REA, an open-source toolkit, lets a coding agent do it. The command bla harden . covers web and desktop apps, Python, databases, AI model files and compiled programs. It has three layers. Poison plants fake clues, Exhaust pads files past the limits of the agent’s analysis tool, and Hallucinate adds decoys, clean-looking fake code the agent may rebuild instead of the real logic.

The author’s tests used REA only, on small test files. BLA padded a desktop-app archive from 607 bytes to 16.9 MiB (a MiB is about a megabyte), mostly in the file index, which REA reads only up to 16 MiB by default, so REA stopped with an error. Across 13 file types, REA hit an error or a limit on 10. BLA’s README makes no claim about Ghidra, Frida or other tools. JavaScript Obfuscator, a code-scrambling tool, rates its free version “fully vulnerable” to AI analysis.

The project says its 14 benchmark samples ran identically after hardening. Its disclaimer says to “protect your own software” and admits no protection inside shipped software can guarantee secrecy against “a dedicated human analyst with unlimited time”. The same decoys could slow an AI that checks malware. REA’s documentation shows how to raise its 16 MiB limit to 256 MiB.

Read More: REA Gives Your Coding Agent a Reverse-Engineering Toolkit, the attack side that BLA is built to resist

Sources:

Disclaimer: For information only. Accuracy or completeness not guaranteed. Illegal use prohibited. Not professional advice or solicitation. Read more: /terms-of-service

Reuse

GNU GENERAL PUBLIC LICENSE v3.0(View License)

Citation

BibTeX citation:
@misc{kabui2026,
  author = {{Kabui, Charles}},
  title = {BLA: {A} {Free} {Tool} {That} {Plants} {Decoys} to {Slow}
    {AI} {Agents} {That} {Reverse} {Engineer} {Your} {Software}},
  date = {2026-11-03},
  url = {https://toknow.ai/posts/bla-blind-llm-agents-defense-against-ai-reverse-engineering/},
  langid = {en-GB}
}
For attribution, please cite this work as:
Kabui, Charles. 2026. “BLA: A Free Tool That Plants Decoys to Slow AI Agents That Reverse Engineer Your Software.” https://toknow.ai/posts/bla-blind-llm-agents-defense-against-ai-reverse-engineering/.