The Protective Shell for Your AI Agents
ClawShell is an AI cost control and security platform built for the OpenClaw ecosystem. Track, budget, and optimize your AI agent spending while scanning for security vulnerabilities.
- Cost Tracking - Real-time spend monitoring per user, agent, model
- Budget Enforcement - Block, alert, or auto-downgrade when limits hit
- Smart Routing - Route to cheaper models based on rules
- Rate Limiting - Per-key rate limits with sliding windows
- Real-time Dashboard - Live updates via Supabase Realtime
- Trust Scoring - 0-100 security scores for AI skills and agents
- Vulnerability Detection - Find secrets, injection risks, and misconfigurations
- Continuous Monitoring - Automated re-scanning of monitored skills
- Compliance Reports - SOC2, ISO27001, and custom frameworks
- Set up infrastructure (Vercel, Supabase, Upstash)
- Copy
.env.exampleto.env.localand fill in credentials - Run database migrations in Supabase SQL Editor
- Deploy:
./scripts/deploy.sh
{
"llm": {
"provider": "anthropic",
"baseUrl": "https://your-clawshell.vercel.app",
"headers": {
"x-clawshell-api-key": "cs_your_key_here"
}
}
}- Product: ClawShell (one word, capital C capital S)
- Security sub-product: ClawShell Scan
- Package:
clawshell,clawshell-scan - Tagline: "The protective shell for your Claw"
MIT