agent-swarm-issue

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 analyzes GitHub issue bodies, comments, referenced PRs and external issue content via gh/GitHub API (e.g., "gh issue view ... --json body,comments" and issue comment commands like "$swarm analyze") and then uses that user-generated, potentially untrusted content to drive automated actions (decomposition, agent assignment, edits, closures), which can enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill repeatedly runs "npx ruv-swarm" (which will fetch and execute the ruv-swarm npm package at runtime) and includes a GitHub Action reference "ruvnet/swarm-action@v1" (which fetches and runs remote action code); both are external runtime dependencies that execute remote code and are relied on for core functionality.