org-ci-health
Warn
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run a local bash script located at
~/.claude/scripts/org-ci-health.shwith a shell command. - [PROMPT_INJECTION]: The user-supplied
org-nameargument is directly interpolated into a bash execution string (~/.claude/scripts/org-ci-health.sh <org-name> > /tmp/org-ci-health.json) without any sanitization or validation logic, which allows for command injection via shell metacharacters. - [COMMAND_EXECUTION]: In Phase 5, the skill performs automated code changes using git worktrees and creates pull requests on GitHub, which involves write access to the local filesystem and remote repository management.
Audit Metadata