proposal-automation

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

• Third-party content exposure detected (high risk: 0.90). The skill's Phase 2.5 competitive analysis and the referenced research methodology (references/competitive-analysis-framework.md and SKILL.md Phase 2.5) explicitly require fetching and ingesting open public sources (G2, Gartner, Google results, YouTube demos, LinkedIn/Twitter/X, Reddit, public websites) and those findings are used to drive Phase 3–4 architecture and autonomous build decisions, so untrusted third‑party content can materially influence agent behavior.