what
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill identifies and ranks technical terminology by parsing local transcript files (.jsonl) and workspace indices (CLAUDE.md, knowledge-index.md). This analysis is performed locally and does not involve network communication or external data transmission.
- [SAFE]: The Python processing script
what_concepts.pyuses standard libraries to extract terms. It incorporates sanitization logic to strip code blocks, URLs, and file paths from processed text, mitigating risks associated with indirect prompt injection. - [SAFE]: The skill includes automated testing and mutation proofs (
mutate.sh,test_what_concepts.py) to verify the reliability of its ranking algorithms. These development tools operate strictly within the local skill environment.
Audit Metadata