health-audit

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly runs "claude plugin list --json" to retrieve plugin metadata from configured marketplaces and parses plugin names/descriptions/keywords (Step 2), which are untrusted/user-generated third‑party contents that the agent reads and uses to decide/add/remove plugins (Step 4/Step 6), allowing indirect prompt injection to influence actions.