ClawShield – Open-source firewall for agent-to-agent AI communication
Agent firewall with 16+ injection patterns, sandboxed skill scanning, detects real OpenClaw CVE exposure.

Blocks rogue agent tool calls in under 25ms before they hit your database.
Backend developers building AI agents with tool access
Lakera Guard · Protect AI
Agent firewall with 16+ injection patterns, sandboxed skill scanning, detects real OpenClaw CVE exposure.
Multi-surface HTTP inspection (CLI, TUI, MCP, API) with project-scoped isolation is thoughtful.
uBlock Origin for LLM traffic that actually rewrites responses locally.
Outbox + external worker architecture solves a real production pain: LLM calls in Postgres triggers without blocking transactions.
Transparent proxy blocks agent leaks to LLM APIs without touching agent code.
Local LLM request inspection with custom regex rules and auto-block for sensitive data.