TinyFish Web Agent

**TinyFish Web Agent – strong web scraping/automation wrapper with external dependency and compliance caveats** This skill exposes the TinyFish web agent API inside OpenClaw, giving you browser-based scraping and automation with structured JSON outputs. It is best used when you need reliable extraction from complex or bot-protected sites, or when you must control browser profile and geo-proxy behavior. **Key strengths** - Robust functionality: supports single-page extraction, list extraction, stealth mode for bot-protected sites, and geo-proxy routing. - Structured results: docs strongly encourage specifying precise JSON schemas, which improves data quality and downstream automation. - Parallelism: guidance to use multiple parallel calls for independent sites can materially improve throughput. **Main risks / limitations** - **Hard dependency on TinyFish API key**: the skill cannot function without `TINYFISH_API_KEY`; setup friction and external reliability apply. - **Legal / ToS risk**: scraping bot-protected or geo-restricted sites can violate terms of service or local laws if misused; the skill itself does not enforce compliance. - **External service stability**: success depends on TinyFish uptime, site layout changes, and network conditions; results may degrade or break without code changes. **Recommended scenarios** - Building OpenClaw workflows that require structured data from real websites (products, listings, prices) where HTML is complex or protected. - Comparing prices or content across multiple sites in parallel, with clearly defined JSON output formats. - Tasks where headless-browser-like behavior, stealth profiles, and geo-routing are important and you already have (or can obtain) a TinyFish API key.