powercat-overflow

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow ingests outsider-authored free text because it loads solution.xml and recursively reads all Workflows/*.json action definitions from a user-uploaded Power Automate solution ZIP (whose internal JSON can contain arbitrary text that the LLM parses at runtime).

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly performs a runtime fetch of the canonical source list at https://raw.githubusercontent.com/microsoft/power-cat-skills/refs/heads/main/Common/PowerCAT%20OverFlow/sources.md and then uses the fetched content to restrict which guideline URLs may be cited, so this external URL directly controls agent behavior.