anysearch

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt requires the agent to extract, handle and potentially persist API keys (including writing them to .env and using them in Authorization headers or via a --api_key flag), which forces the LLM to observe and may cause it to output secret values verbatim (e.g., when asking to save or returning the key), creating an exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.85). At runtime, the skill’s search/batch_search/extract commands call the AnySearch API (https://api.anysearch.com/mcp) and then print the returned result.content[0].text (or full JSON) to stdout, which the agent can ingest as LLM context; that returned text is derived from outsider-authored web/social content.