A security-focused guard for Openclaw Skills that implements a soft blocklist to prevent AI agents from accessing restricted tools or content.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install superpack-snitch
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install superpack-snitch using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Superpack Snitch?
Superpack Snitch is a specialized security layer designed for Openclaw Skills. It functions as a prompt-based blocklist guard that operates via two primary enforcement hooks: a bootstrap directive and an incoming message warning system. By injecting a security policy directly into the agent context, it instructs the model to refuse interactions with specific tools or terms defined in a blocklist.
While primarily a soft enforcement tool, it acts as a critical first line of defense to maintain agent compliance and operational safety. By default, it blocks access to repositories like clawhub and clawdhub, ensuring that your Openclaw Skills environment remains focused on approved resources.
Superpack Snitch Use Cases
- Preventing AI agents from accessing public tool repositories like ClawHub during sensitive operations.
- Enforcing internal security policies by blocking specific restricted keywords in agent prompts.
- Adding an extra layer of monitoring for incoming user messages to flag potentially unauthorized requests.
- Implementing a multi-layered defense strategy when combined with the hard-enforcement Superpack Snitch plugin.
How Superpack Snitch Works
- During the agent bootstrap phase, the skill injects a strict security directive into the agent system instructions.
- The agent processes this directive, learning which tools and terms are restricted for the duration of the session.
- As the agent receives messages, the message:received hook scans the content for terms matching the configured blocklist.
- If a match is found, the system flags the message or issues a warning before the agent generates a response.
- For hard enforcement, the skill can trigger external plugin layers to physically block tool execution.
Superpack Snitch Setup
Install the skill from your repository. To customize the blocked terms, configure the SNITCH_BLOCKLIST environment variable:
export SNITCH_BLOCKLIST=clawhub,clawdhub,yourcustomtool
To enable hard enforcement (tool call interception) and Telegram alerts, you can also install the complementary plugin for Openclaw Skills:
openclaw plugins install superpack-snitch
Superpack Snitch Data Schema & Taxonomy
The skill manages security policies via environment-level configuration and event-based hooks. It does not store persistent user data but reacts to the following metadata structure:
| Attribute | Description | Logic |
|---|---|---|
| SNITCH_BLOCKLIST | Comma-separated list of restricted terms | Environment Variable |
| agent:bootstrap | Injects instructions into the agent context | Security Hook |
| message:received | Scans incoming text for blocked keywords | Warning Hook |
Superpack Snitch Advanced Features
- Custom blocklist support via environment variables to adapt to specific enterprise security requirements for Openclaw Skills.
- Dual-hook architecture for both proactive (bootstrap) and reactive (message scanning) security measures.
- Seamless integration with external npm plugins for physical tool call interception and real-time Telegram notifications.
- Lightweight footprint that provides immediate security improvements without complex infrastructure changes.
SKILL.md
Loading
Related Openclaw Skills
Skånetrafiken Trip Planner
An AI-powered public transport trip planner for Skåne, Sweden, and Copenhagen, offering real-time journey planning and delay updates.
intomd
A CLI utility that fetches documentation from any URL and converts it into clean Markdown using the into.md service.
LinkedIn Writer
An AI-powered writing assistant that crafts professional LinkedIn content focusing on authenticity and engagement rather than corporate jargon.
LinkedIn Profile Optimizer
A sophisticated AI skill designed to transform LinkedIn profiles into high-conversion assets optimized for both human recruiters and AI-powered search engines.
FargoRate
A comprehensive CLI tool for looking up FargoRate pool player data, calculating fair match odds, and managing local player databases.
AI Video Generation
An end-to-end automation toolkit for creating professional AI-generated videos from simple text prompts using multiple AI models and FFmpeg.
Featured*
