spawn

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

• Third-party content exposure detected (high risk: 0.70). The skill's Implementation Steps explicitly require using "WebSearch or WebFetch" to "research official docs" and "find 10+ authoritative URLs" (and asks for 10–15 canonical resources), which entails fetching open/public third‑party web content as part of its generation workflow and could surface untrusted user-controlled pages.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly requires runtime "Research official docs (WebSearch or WebFetch)" and to include authoritative links (e.g., https://docs.claude.com/en/docs/claude-code/sub-agents) which would be fetched and incorporated into/generated prompts, so that external content can directly influence the agent's instructions.