clawsec-suite

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

• Third-party content exposure detected (high risk: 0.80). The hook and scripts explicitly fetch and parse remote community-driven indexes/feeds (e.g., loadRemoteFeed in hooks/clawsec-advisory-guardian/lib/feed.mjs pulls https://clawsec.prompt.security/advisories/feed.json and scripts/discover_skill_catalog.mjs fetches https://clawsec.prompt.security/skills/index.json), treat those community-sourced advisories as untrusted/user-generated input, and use them to drive alerts and approval-gated install/remove decisions—so third-party content is read and can materially change agent behavior.