lab-homework

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

• Third-party content exposure detected (high risk: 0.90). The skill fetches the public curriculum manifest JSON (including outsider-authored summary_md text) from https://agency-lab.glebkalinin.com/api/curriculum/<cohort>.json at runtime and then grounds the generated homework in that summary_md.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches the curriculum manifest at runtime from https://agency-lab.glebkalinin.com/api/curriculum/.json (and optionally https://agency-lab.glebkalinin.com/api/curriculum//.json), and that fetched JSON (summary_md, meetings[]) directly grounds and controls the agent's generated assignment prompts, so these endpoints are runtime dependencies that control prompts.