anysearch

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

• Third-party content exposure detected (high risk: 1.00). This skill explicitly calls the AnySearch API (https://api.anysearch.com/mcp) and includes an extract command to fetch arbitrary web pages and a mandatory list_domains step the agent "MUST" read/obey (see SKILL.md Decision Flow and scripts/anysearch_cli.*), so it ingests untrusted third‑party web content that can directly influence subsequent tool use and decisions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill's CLI scripts call the external JSON‑RPC endpoint https://api.anysearch.com/mcp at runtime (ENDPOINT in multiple scripts) and the README/SKILL.md explicitly requires calling list_domains and "read query_format from result" so responses from that URL directly control how the agent constructs prompts/queries.