tavily-search

Tavily AI search API integration for OpenClaw. Provides web search functionality with AI-powered summarization optimized for RAG and question answering. Use...

This package appears to implement a legitimate Tavily search client, but there are several packaging/information mismatches you should resolve before installing: - The code requires a TAVILY_API_KEY environment variable but the registry metadata does not declare this; expect the CLI to fail if you don't set TAVILY_API_KEY. Ask the publisher to add the required env var/primary credential to the metadata. - The SKILL.md example references a non-existent script (tavily_search.py vs the provided tavily_cli.py). Confirm which entrypoint to run. - The code sends your queries and the API key to https://api.tavily.com as JSON. If you will send sensitive queries, be aware they leave your environment and go to that external service; review Tavily's privacy and retention policy before use. - Confirm the source/publisher (owner ID looks opaque and homepage is missing). If you need a higher trust level, request a verifiable homepage or repository and a signed release. - Ensure the runtime has the requests library available; the package does not declare dependencies in metadata. If you accept those caveats and set TAVILY_API_KEY securely, the code itself does not contain obvious exfiltration or obfuscated behavior. If you need stronger assurance, ask the publisher to correct metadata, provide a canonical repo link, and sign the release.

Type: OpenClaw Skill Name: tavily-search-quant Version: 1.0.0 The skill bundle provides a legitimate integration with the Tavily Search API for web searching, Q&A, and content extraction. The implementation in `scripts/tavily_api.py` and `scripts/tavily_cli.py` uses standard Python libraries (requests, argparse) to interact with the official Tavily endpoint (api.tavily.com) and correctly handles API keys via environment variables. No evidence of malicious execution, data exfiltration, or prompt injection was found.