joyco-logs

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md requires runtime curl/WebFetch of public JOYCO index pages and then fetching matched /logs/NN-slug.md or /toolbox/slug.md markdown, which is outsider-authored free text from public web content that can be ingested into the agent’s LLM context.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches remote markdown at runtime (e.g., curl -s https://hub.joyco.studio/logs.md) and tells the agent to read and apply those documents to guide its behavior, so the hub URLs directly control agent instructions.