dev-workflow-triage

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and parses user-generated GitHub issues via "gh issue list --repo SonicGarden/dev-workflow-issues" (Step 2) and then reads each issue body / Finding (Step 3 Parse body) to drive judgments, edits, and downstream Skill(verify-diff)/Skill(skill-review)/Skill(publicity-review) actions, so untrusted third‑party issue content can materially influence tool behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill runs git fetch origin 'refs/heads/triage-*' and then switches the run branch (and also uses gh to read the repo) at runtime against the repository origin (e.g. https://github.com/SonicGarden/dev-workflow-issues), and the fetched repository files (skills//SKILL.md) are then read and used as authoritative skill instructions that directly control agent behavior — a remote branch on the origin can therefore change the agent's prompts/instructions during execution.