skill-review

Pass

Audited by Gen Agent Trust Hub on Apr 15, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands to inventory skill files and check git history. These operations are restricted to retrieving local file metadata (stat, date) and viewing the commit log (git log).
  • [DATA_EXFILTRATION]: No network operations or external data transfers were identified. The skill only interacts with local files and the local git repository.
  • [PROMPT_INJECTION]: The instructions focus entirely on maintenance workflows and do not contain any patterns aimed at bypassing safety guidelines or overriding agent constraints.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface by reading other skill files (skills/*/SKILL.md) and shared lesson files (shared/gotchas.md). However, these are internal configuration files rather than untrusted external data, posing negligible risk in this context.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 15, 2026, 06:31 AM