autoskill

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md's "Similar Skill Search" explicitly instructs the agent to check external public sources (e.g., "browse https://skills.sh/ or its leaderboard" and "run npx skills find ") and to present external matches as part of the create/improve/merge decision flow, meaning untrusted third-party content is fetched and can materially influence agent decisions.