write-release-notes

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

• Third-party content exposure detected (high risk: 1.00). The skill's workflow and bundled script (SKILL.md "Fetch Release Data" step and scripts/fetch-release-data.sh) explicitly pull live release metadata and each linked PR's title/body from GitHub — untrusted, user-generated web content that the agent is required to read and synthesize into release notes, so those external texts could influence its outputs or behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill runs GitHub API calls at runtime (e.g. gh api "repos/$REPO/releases/tags/$VERSION" — i.e. https://api.github.com/repos///releases/tags/ and gh pr view for PR details) to fetch release and PR bodies that are injected into the agent's context and thus directly control the generated prompts/output.