create-command

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill's example commands explicitly fetch and ingest user-generated content from GitHub (e.g., !gh pr view $1, !gh pr diff $1, gh repo view --json ...) and then instruct the agent to read and analyze those PRs/issues, exposing the agent to untrusted third-party content.