fork-intelligence

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 analyzes public, user-generated GitHub content (e.g., via the gh api calls in the Pipeline steps such as "gh api 'repos/$UPSTREAM/forks' --paginate", "gh api 'repos/FORK_OWNER/REPO/branches'", and "gh api 'repos/$UPSTREAM/pulls?state=all'") and requires the agent to read commit messages, branch names, PRs, tags, and issues as inputs that materially influence fork-ranking and follow-up actions, so untrusted third-party content can indirectly inject instructions.