thoughtful-commitment

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly ingests user-generated PR titles/bodies from GitHub (e.g., the examples and context-gathering commands like gh pr view / gh pr list in SKILL.md and CARD.yml) and uses those PR descriptions as context for generating commit messages, so it reads untrusted third-party content as part of its workflow.