skill-forge

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

• Third-party content exposure detected (high risk: 0.90). The skill's Phase 1A "Research (URL-based)" flow (and front description "For URLs: performs deep doc research (README, llms.txt, API references)") explicitly instructs the agent to crawl and ingest arbitrary public GitHub repos and documentation URLs, i.e., untrusted third‑party/user‑generated content that the agent must read and that will directly shape generated skills and actions.

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

• Potentially malicious external URL detected (high risk: 0.90). This skill explicitly crawls user-provided URLs at runtime (e.g., https://github.com/resendlabs/resend-node and related docs like https://resend.com/llms.txt) and injects the fetched README/llms.txt/API content into the skill-generation process, so those external URLs directly control the prompts/instructions the agent uses.