txtskills-llms-to-agent-skills

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and parses llms.txt from arbitrary public documentation URLs (see SKILL.md agent execution workflow and apps/web/app/api/convert/route.ts which calls fetchLlmsTxt to retrieve and parse remote llms.txt content) and then uses that content to generate/publish installable agent skills, so untrusted third‑party content is read and can directly influence generated skill behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The convert API fetches user-supplied documentation URLs at runtime (e.g., https://docs.example.com/llms.txt or /llms.txt used via npx txtskills@latest convert) and then directly parses that fetched content to generate/publish Agent Skills, so remote llms.txt content can directly control prompts/instructions used by agents.