jq

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md Agent-Specific Notes explicitly show fetching arbitrary URLs with curl (e.g., "curl -s url | jq '.data'"), which instructs the agent to ingest and parse untrusted public/web API JSON as part of its workflow and could therefore enable indirect prompt injection.