maintenance

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform system-level maintenance operations. This includes deleting old log files using find -delete in the .claude/logs/ directory and truncating state files using tail and mv within the .claude/state/ directory. These commands are localized to the project's metadata folders.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from files such as Plans.md and session-log.md, which are updated during normal agent operations and may contain user-supplied content. The agent reads these files to identify tasks for archiving or logs for rotation. While the skill lacks explicit boundary markers for these inputs, the logic is limited to structural file organization rather than executing instructions contained within the text.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 07:59 AM
Security Audit — agent-trust-hub — maintenance