gh-issue-report

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

• Third-party content exposure detected (high risk: 1.00). The skill fetches and reads untrusted, user-authored repository content (e.g., CONTRIBUTING.md, .github/ISSUE_TEMPLATE files, and source files) via gh api calls like 'repos//contents/...' as described in Phase 3 and Phase 4, and it uses that content to extract conventions, labels, and to compose/file issues—so third-party content can influence agent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill issues runtime gh API calls to fetch repository files (for example: https://api.github.com/repos//contents/.github/ISSUE_TEMPLATE/), and those fetched templates/contributing guides are injected into the agent's issue-drafting logic (i.e., directly control prompts), so this is a runtime external dependency that influences agent instructions.