Clawsec.bot

Clawsec.bot

clawsec.bot

AI agent security plugin

Install with one command and block dangerous AI agent actions in under 5ms.

FreeLinuxmacOSWindowsAI Agents
Visit website for Clawsec.bot (opens in a new tab)

Overview

Clawsec is an open-source security plugin for OpenClaw that intercepts tool calls from AI agents and blocks dangerous actions in under 5ms. It protects against file deletion, secret leakage, data exfiltration, unauthorized purchases, privilege escalation, and prompt injection. Designed for developers running OpenClaw agents, it installs with one command and requires zero configuration.

How it works

Clawsec works by intercepting every tool call an OpenClaw agent makes, including shell commands, HTTP requests, and file writes. It uses a rules engine to score risk against built-in rulesets like destructive-commands, secrets, and exfiltration. Based on the matched rule's configured action in clawsec.yaml, it can block the action, require human approval, or allow and log it with a full audit trail.

Key capabilities

Clawsec catches destructive commands like rm -rf and DROP TABLE, secret exposure of API keys and tokens, data exfiltration via curl to external endpoints, unauthorized purchases through payment APIs, privilege escalation with sudo and chmod 777, and prompt injection attempts. Every rule is configurable and extensible through YAML, and new rules ship with each release.

Who it is for

Clawsec is for developers and teams running OpenClaw AI agents who need to enforce security guardrails. It is especially useful for those deploying agents with root access who want to prevent accidental or malicious actions. The plugin is open source and MIT licensed, making it suitable for individual developers and enterprises alike.

Pricing and limits

Clawsec is free and open source under the MIT license. Installation is a single command: openclaw plugins install clawsec. It auto-generates a clawsec.yaml configuration file, and users can extend it with built-in rulesets like aws-security and pii-protection. The plugin works with any agent framework and adds under 5ms latency per tool call.

Key features

  • Ultra-low latency

    Blocks dangerous actions in under 5ms

  • Comprehensive rulesets

    Built-in rulesets for destructive commands, secret exposure, data exfiltration, unauthorized purchases, privilege escalation, and prompt injection

  • Configurable actions

    Extensible YAML rules with actions: block, confirm, or allow

  • Quick setup

    One command install: openclaw plugins install clawsec

  • Open source

    MIT licensed and open source

Use cases

  • Block destructive commands

    Prevent AI agents from executing destructive commands like rm -rf or DROP TABLE

  • Prevent secret exposure

    Stop leakage of API keys, tokens, and credentials in agent outputs

  • Prevent data exfiltration

    Block unauthorized uploads to external endpoints

  • Control unauthorized purchases

    Require human approval for financial transactions and subscription changes

  • Mitigate prompt injection

    Detect and block prompt injection attempts that try to override system instructions

Pricing & limits

From the official site · checked 10/10/2026
Open SourceFree
See current pricing on the official site

Frequently asked questions about Clawsec.bot

What is Clawsec?
Clawsec is an open-source security plugin for OpenClaw that intercepts tool calls from AI agents and blocks dangerous actions like file deletion, secret leakage, and unauthorized purchases.
Who is Clawsec for?
It is for developers and teams running OpenClaw AI agents who need to enforce security guardrails on agent actions.
Is Clawsec free?
Yes, Clawsec is open source and MIT licensed. It is free to use and install via the command 'openclaw plugins install clawsec'.
Which platforms does Clawsec support?
Clawsec works with OpenClaw, an AI agent framework. It intercepts tool calls for shell commands, HTTP requests, and file writes.