daigest

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

• Third-party content exposure detected (high risk: 0.80). This skill's SKILL.md shows creating feeds with arbitrary "rss" or "url" sources and calling POST /feeds/{id}/sync and GET /feeds/{id}?since=... to fetch the generated markdown digest that the agent is explicitly told to "use," so untrusted public web/social content is ingested and can influence agent decisions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill makes runtime calls to the Daigest API (e.g., https://daige.st/api/v1 and its OpenAPI spec https://daige.st/api/v1/openapi.json) and the service returns "memory" and feed content that directly supply AI instructions/prompts which the agent uses, so this external URL controls agent prompts at runtime.