manage-knowhow-capture

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

• Third-party content exposure detected (high risk: 0.85). For reference type, the workflow fetches arbitrary public web content at runtime from --source <url> via curl, then extracts readable text into the LLM context for summarization (indirect prompt injection risk from outsider-authored page content).

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

• Potentially malicious external URL detected (high risk: 0.90). This skill explicitly fetches --source URLs at runtime via Bash curl and injects the fetched content into generated references (example: https://docs.stripe.com/api), so remote content can directly control the agent's output.