reference-checker

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

• Third-party content exposure detected (low risk: 0.10). The workflow reads arbitrary web pages and scholarly sites to verify references and bibliography entries, which involves processing outsider-authored content.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs connecting to a third-party marketplace, plugin repository, and remote MCP server (https://claude.ai/directory/hedwigai, un-seen/hedwigai-skills, https://mcp.hedwigai.com/) with unknown and unverified provenance, representing a high-risk external operational dependency.