memento-reflect
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a reflective learning protocol that involves analyzing conversation history to determine task success or failure. This analysis is restricted to the current session context.
- [DATA_EXFILTRATION]: File operations are limited to reading and writing a specific metrics file located in
~/.claude/memento/metrics.jsonor the local project directory. No access to sensitive system paths (e.g., SSH keys, AWS credentials) or unauthorized network communications were detected. - [COMMAND_EXECUTION]: The skill mentions external command patterns like
/memento createand/memento optimize. These are presented as user-facing suggestions for manual or guided execution and do not involve hidden or arbitrary shell command injection.
Audit Metadata