weekly-review

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

• Third-party content exposure detected (high risk: 0.80). The skill runs a GitHub CLI command (gh search commits ...) to fetch commit data from GitHub — which is user-generated, potentially public third‑party content — and the agent is expected to read and summarize those commit messages as part of its workflow, exposing it to indirect prompt injection risk.