work-summary
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows security best practices by explicitly instructing the agent to remain read-only towards local history databases and files, preventing accidental data modification or deletion.
- [SAFE]: Data privacy is addressed by requiring all history content to remain local. The instructions specifically forbid sending data to external services and warn against staging or committing generated reports to Git, as they may contain sensitive snippets from private projects.
- [SAFE]: The skill implements clear boundary rules for its optional Korean prose enhancement feature, stating it must not install, fetch, or require external tools if they are not already present in the environment.
- [SAFE]: The command execution patterns involve standard local file reading (e.g.,
jq,sqlite3, and file globbing) targeting known local application directories (~/.claude,~/.codex, etc.). No suspicious network operations or privilege escalation attempts were found. - [SAFE]: The skill provides clear instructions for handling 'empty' ranges by reporting 'no activity' rather than inventing data, which prevents hallucination-based misinformation.
Audit Metadata