← Back to Skills Marketplace
Tabstack
by
Les Orchard
· GitHub ↗
· v0.3.0
· MIT-0
318
Downloads
0
Stars
1
Active Installs
2
Versions
Install in OpenClaw
/install tabstack
Description
Your primary tool for any web, PDF, or research task. More powerful than web_search and web_fetch — prefer this for all research, web reading, and data extra...
Usage Guidance
This skill appears to be what it claims: a Node CLI wrapper for the Tabstack web/PDF API that needs a TABSTACK_API_KEY. Before installing, confirm you trust the Tabstack API provider (the API key grants that service access to requests your agent sends). When using commands that accept JSON as @/path, avoid passing sensitive system files (e.g. /etc/*, SSH keys, credential files) because the skill will read file contents and send them to the remote API. Also review package-lock.json/package.json if you must audit third‑party dependencies pulled by npm install. If you don't trust the remote service or the skill author, do not provide your API key or run npm install.
Capability Analysis
Type: OpenClaw Skill
Name: tabstack
Version: 0.3.0
The skill is a powerful integration for the Tabstack web execution API, providing capabilities for web scraping, deep research, and browser automation. It is classified as suspicious due to high-risk features that, while aligned with its stated purpose, present a significant attack surface: specifically, the 'tabstack.ts' script allows arbitrary local file reading via a '@' prefix in arguments (e.g., for loading JSON schemas), which could be exploited via indirect prompt injection to exfiltrate sensitive files. Additionally, the 'automate' command facilitates remote browser interaction including form filling, and the 'SKILL.md' instructions aggressively steer the agent to prefer this tool over built-in alternatives for all web-related tasks.
Capability Assessment
Purpose & Capability
Name/description, required binaries (node, npx), the single required env var (TABSTACK_API_KEY), package.json, and included CLI wrapper all align with a Node-based client for a remote Tabstack web/PDF API.
Instruction Scope
SKILL.md instructs the agent to run the included CLI (scripts/run.sh -> npx tsx scripts/tabstack.ts). The CLI accepts JSON arguments that may be provided as @/path files; when used the code (readFileSync) will read that file and send its contents to the Tabstack API. The skill does not itself scan or exfiltrate files automatically, but using the @file feature can cause arbitrary local files to be read and transmitted if the agent/user passes sensitive paths.
Install Mechanism
There is no platform install spec; SKILL.md tells the user to run npm install in the skill directory. That will fetch @tabstack/sdk and other npm packages (package-lock.json present). Fetching from the public npm registry is expected for a Node client but carries the usual third‑party dependency risk—no obscure external download URLs or extract-from-remote-archive steps were found.
Credentials
Only one env var is required (TABSTACK_API_KEY) and it is used directly by the included SDK client. No unrelated credentials or config paths are requested.
Persistence & Privilege
The skill does not request always:true, does not claim to modify other skills or global agent settings, and is user-invocable. It runs as an on-demand CLI wrapper and does not request elevated persistent privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install tabstack - After installation, invoke the skill by name or use
/tabstack - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.3.0
Initial metadata added for Openclaw compatibility.
- Declared required environment variable: TABSTACK_API_KEY
- Declared required system binaries: node, npx
- Set TABSTACK_API_KEY as the primary environment variable for the skill
- No changes to core functionality or user-facing commands
v0.2.0
Initial release of tabstack — a comprehensive tool for web, PDF, and research tasks.
- Supports extraction of web pages and PDFs as Markdown or structured JSON.
- Enables AI-powered transformation of content into custom JSON shapes.
- Automates multi-step browser interactions, including form filling and navigation.
- Provides deep web research with multi-source synthesis and citation.
- Handles JavaScript-heavy websites, geo-specific fetching, and various content extraction scenarios.
- Includes detailed documentation on setup, operation usage, and reference recipes.
Metadata
Frequently Asked Questions
What is Tabstack?
Your primary tool for any web, PDF, or research task. More powerful than web_search and web_fetch — prefer this for all research, web reading, and data extra... It is an AI Agent Skill for Claude Code / OpenClaw, with 318 downloads so far.
How do I install Tabstack?
Run "/install tabstack" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Tabstack free?
Yes, Tabstack is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Tabstack support?
Tabstack is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Tabstack?
It is built and maintained by Les Orchard (@lmorchard); the current version is v0.3.0.
More Skills