← Back to Skills Marketplace
TinyFish Web Agent
by
Simantak Dabhade
· GitHub ↗
· v1.0.3
· MIT-0
3003
Downloads
3
Stars
4
Active Installs
4
Versions
Install in OpenClaw
/install tinyfish-web-agent
Description
Use TinyFish web agent to extract/scrape websites, extract data, and automate browser actions using natural language. Use when you need to extract/scrape dat...
Usage Guidance
This skill appears to be a front-end for a remote scraping service (TinyFish) and its instructions require installing an npm CLI and providing an API key (TINYFISH_API_KEY), but the registry metadata did not declare those requirements. Before installing or using it: (1) verify the publisher and the npm package (@tiny-fish/cli) on the npm registry — inspect its source repo and recent releases; (2) confirm the service endpoint (agent.tinyfish.ai) and read its privacy/data retention policy (scraped page contents will be transmitted to that service); (3) do not use your personal or sensitive credentials/pages with the skill unless you trust the service; (4) ask the skill author to correct registry metadata to list TINYFISH_API_KEY as a required credential and provide a homepage/source repo; (5) consider installing/testing the CLI in an isolated environment or container first. The primary inconsistency is the undisclosed credential requirement and implied remote upload of scraped content — that makes this skill suspicious, not conclusively malicious.
Capability Analysis
Type: OpenClaw Skill
Name: tinyfish-web-agent
Version: 1.0.3
The skill bundle provides instructions for an AI agent to use the TinyFish CLI for web searching, content extraction, and browser automation. The SKILL.md file contains legitimate documentation, usage patterns, and command examples (e.g., search, fetch, agent run) that align with the stated purpose of a web agent. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found; the tool uses standard authentication via API keys and interacts with the tinyfish.ai domain.
Capability Tags
Capability Assessment
Purpose & Capability
The skill's name/description describe a web-scraping/browser-automation agent and the SKILL.md consistently documents tools (search, fetch, agent, browser) that fit that purpose. However, the SKILL.md expects use of a remote TinyFish service and an API key, which the registry metadata does not declare.
Instruction Scope
The instructions stay within the claimed domain (search/fetch/agent/browser automation). They explain command usage, flags, batching, and streaming outputs; they do not instruct reading unrelated system files. They do instruct sending URLs/content to the remote TinyFish service via the CLI.
Install Mechanism
There is no formal install spec in the registry, but the SKILL.md recommends installing @tiny-fish/cli via `npm install -g`. Installing a global npm package will download and run third-party code — expected for a CLI but a moderate-risk action that the metadata did not surface.
Credentials
The SKILL.md explicitly references an env var (TINYFISH_API_KEY) and an auth flow (tinyfish auth login), but the registry metadata shows no required environment variables or primary credential. This mismatch hides a required credential and the fact that scraped content will be sent to an external service.
Persistence & Privilege
The skill does not request always:true and does not declare modifications to other skills or system-wide settings. Autonomous invocation is allowed (platform default) but not by itself a red flag here.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install tinyfish-web-agent - After installation, invoke the skill by name or use
/tinyfish-web-agent - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.3
- Switched from shell-based API instructions to CLI-based usage with `tinyfish` commands.
- Comprehensive new documentation on four TinyFish tools: search, fetch, agent, and browser.
- Added clear guidance on tool selection, escalation workflow, and parallelization best practices.
- Provides detailed usage examples, flags, and output structure for all CLI commands.
- Removed the deprecated `scripts/extract.sh` script file.
v1.0.2
- Switched API environment variable from MINO_API_KEY to TINYFISH_API_KEY.
- Updated all examples from Python to curl/bash for easier integration.
- Added strict pre-flight key verification instructions before any API call, with clear user messaging if the key is missing.
- Specified new homepage and environment key requirements in metadata.
- Removed extract.py script and added extract.sh to reflect the transition to shell usage.
v1.0.1
Summary: Added best practices for specifying JSON output and running parallel extractions.
- Introduced a "Best Practices" section emphasizing explicit JSON format and parallel API calls for independent sites.
- Updated example extraction prompts to specify desired JSON output structure.
- Added guidance and code examples demonstrating parallel extraction versus single combined prompts.
- No code or feature changes; documentation improvements
v1.0.0
- Updated skill.
Metadata
Frequently Asked Questions
What is TinyFish Web Agent?
Use TinyFish web agent to extract/scrape websites, extract data, and automate browser actions using natural language. Use when you need to extract/scrape dat... It is an AI Agent Skill for Claude Code / OpenClaw, with 3003 downloads so far.
How do I install TinyFish Web Agent?
Run "/install tinyfish-web-agent" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is TinyFish Web Agent free?
Yes, TinyFish Web Agent is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does TinyFish Web Agent support?
TinyFish Web Agent is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created TinyFish Web Agent?
It is built and maintained by Simantak Dabhade (@simantak-dabhade); the current version is v1.0.3.
More Skills