create-integration

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly requires fetching and reading external API/docs URLs and user-provided research/sample files (see Phase 1: "Read any @-mentioned files. Fetch any documentation or API URLs provided inline" and the "What to provide" rows that accept pasted API/log documentation URLs), so untrusted third-party web content or user-supplied docs could be ingested and influence subsequent subagent actions and tool use.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs the agent to "Fetch any documentation or API URLs provided inline" at runtime (e.g., the example URL https://docs.acme.com/api/v2/events), and those fetched docs are used to build subagent prompts and implementation details, so external content can directly control agent instructions.