sweep-reds
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of natural language instructions in a markdown file. It does not include any scripts, executable binaries, or configuration files that trigger code execution.
- [SAFE]: The instructions establish strong security and operational boundaries, explicitly prohibiting the agent from merging pull requests, closing PRs, weakening test assertions, or accessing credentials without owner authorization. It focuses on isolated reproduction of failures and preparation of fixes for human review.
Audit Metadata