rightcode-skill-updater

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

• Third-party content exposure detected (high risk: 0.80). The skill's check.py and apply.py explicitly fetch and decode public GitHub content via the gh API from third‑party repos (e.g., mattpocock/skills and awesome-skills/code-review-skill) and then read/compare that upstream SKILL.md and supporting files to drive diffs, interactive apply decisions, and to write those files locally, meaning untrusted user-authored repo content is ingested and can influence what the tool does and what agent-visible skills get created.

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

• Potentially malicious external URL detected (high risk: 0.90). The scripts call gh api at runtime to fetch and decode SKILL.md and other files from https://github.com/mattpocock/skills and https://github.com/awesome-skills/code-review-skill, and those fetched SKILL.md contents are written into local skill definitions (thus directly controlling agent prompts/instructions), so these upstream GitHub URLs are a required runtime dependency.