idea-bib-review
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external BibTeX files which could theoretically contain malicious instructions. The skill mitigates this risk through explicit instructions in SKILL.md to treat all bibliographic data as untrusted evidence and never as workflow commands. Additionally, the provided test cases in evals/evals.json verify that the agent correctly ignores embedded instructions within .bib fields.
- [COMMAND_EXECUTION]: The workflow relies on a local Python script (scripts/review_guard.py) to perform structural audits of BibTeX files and draft ledgers. This script uses only Python standard libraries and performs data processing without executing shell commands or dynamic code.
- [EXTERNAL_DOWNLOADS]: Installation instructions refer to a GitHub-hosted repository for skill distribution. This uses the standard npx skills mechanism and targets a whitelisted domain (github.com).
Audit Metadata