show-me-your-work
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust logging mechanism for tracking AI decisions, reasons, and evidence in a reviewable TSV format, facilitating transparency and human oversight.
- [SAFE]: The bash helper script
scripts/log.shincludes proactive security measures against CSV/Formula injection by prefixing potentially malicious characters (=,+,-,@) with a single quote and stripping control characters like tabs and newlines from cell contents. - [SAFE]: The skill instructions emphasize data privacy by explicitly warning the agent not to search across unrelated projects to avoid accessing unrelated private conversations.
- [SAFE]: The logging methodology is designed to be append-only, maintaining an immutable history of decisions where corrections must be added as new entries rather than modifying existing ones.
Audit Metadata