build-feature

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly fetches and parses GitHub issue bodies, project board items, and referenced decision records (via gh commands and reading the decision record path in the issue) and uses that user-generated content to extract the implementation plan and drive branching, task execution, and commits, so untrusted third-party issue content could inject instructions that affect agent behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches and parses GitHub issue content at runtime (e.g., https://github.com/{OWNER}/{REPO}/issues/) and reads the referenced decision record to drive the implementation plan and agent actions, so remote issue/decision content directly controls the agent's instructions.