skill-repo

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

• Potentially malicious external URL detected (high risk: 0.90). The SR-34 LLM review prompt explicitly directs the runner to fetch and read https://raw.githubusercontent.com/netresearch/skill-repo-skill/main/docs/ARCHITECTURE.md at runtime to determine CI workflow behavior, meaning an external URL is fetched during skill runtime and its content directly controls the review instructions.