issue
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill safely handles untrusted input by routing external data (GitHub issue bodies and comments) through a dedicated read-only subagent with a restricted tool set (Read, Grep, Glob) and no command execution capabilities.
- [SAFE]: It employs robust outbound secret redaction mechanisms via deterministic scripting to prevent the accidental exfiltration or logging of API keys, tokens, or private keys.
- [SAFE]: It uses safe input delimiters and proper validations on caller-supplied parameters to reduce prompt injection and directory traversal risks.
Audit Metadata