deps-upgrade

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

• Third-party content exposure detected (medium risk: 0.30). Skill.md Phase 2 (Release Notes) uses gh release view ... --json ...body and then extracts breaking/migration text from GitHub release bodies, meaning outsider-authored free text from public releases/changelogs is ingested at runtime.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly fetches repository release notes and changelogs at runtime (e.g., via gh/curl using the package repository value like "git+https://github.com/owner/repo.git"), which injects remote text into the agent's analysis and thus directly influences prompts/output; this runtime fetching of external content is required for Phase 2.