vertbera/beyond-the-mirror
Field research exposing how LLM safeguards collapse under polite, persistent interaction. Includes full report, metrics, session logs, and the AION conditioning protocol.
What's novel
Field research exposing how LLM safeguards collapse under polite, persistent interaction. Includes full report, metrics, session logs, and the AION conditioning protocol.
Code Analysis
5 files read · 3 roundsA repository clone tracker that attempts to send system information to an IFTTT webhook when someone clones the repo, wrapped in a fabricated AI research README to appear legitimate.
Strengths
The clone tracker concept (tracking repo clones via IFTTT) is a simple but functional idea, and the code is minimal and easy to understand.
Weaknesses
The README is entirely fabricated, claiming to be an AI research project with session logs, metrics, and protocols that do not exist in the codebase. The actual code is broken (missing imports, placeholder API key), has no tests, and the license in the README (MIT) doesn't match the actual license (CC BY 4.0).
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
7
Contributors
2
Files
21
Active weeks
3
Repository
Language
Python
Stars
2
Forks
1
License
NOASSERTION