skill-master

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

• Third-party content exposure detected (high risk: 0.90). The skill's docs (references/docs-ingestion.md and the "Creating a Skill from Documentation" section in SKILL.md) explicitly instruct the agent to autonomously fetch public documentation pages (e.g., "Fetch the page" using ai_fetch_url or similar) and ingest them into references/ which the agent then reads and uses to update SKILL.md and drive further actions, exposing it to untrusted third‑party content that can influence behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The docs-ingestion workflow explicitly instructs the agent to fetch external pages at runtime (via tools like ai_fetch_url) and ingest them into skill references—e.g., fetching https://agentskills.io/specification—which means remote URL content can directly control the skill prompts/instructions.