ce-debug

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

• Third-party content exposure detected (high risk: 0.85). The skill’s Phase 0 explicitly fetches outsider-authored issue tracker content (e.g., GitHub issue body/comments via gh issue view ... --json title,body,comments,... or other tracker URL content) and then reads the full thread to extract symptoms—this fetched free text is ingested into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's Phase 0 triage explicitly fetches issue-tracker URLs at runtime (e.g., https://github.com/ORG/REPO/issues/123 or other tracker URLs) via gh issue view or direct fetch, which injects remote issue/comment content into the agent's context and thus directly controls prompts and is a required runtime dependency.