Anti-Virus for Autonomous Agents
Runtime security that intercepts every command your AI agent runs. Block dangerous actions, redact leaked secrets, and detect prompt injection — before damage is done.
npm i -g @contextfort-ai/openclaw-secure
Bash Guard
Static analysis engine that inspects every shell command before execution.
Powered by the Tirith rule engine, the Bash Guard intercepts every child_process call (spawn, exec, execFile) and runs static analysis checks before the command executes. It catches dangerous patterns like recursive deletes, privilege escalation, data exfiltration via curl/wget, and obfuscated payloads. Commands are blocked before they ever run — the LLM never sees the output.
Blocks rm -rf /, chmod 777, and other destructive commands
Detects data exfiltration via curl, wget, nc, and DNS tunneling
Catches obfuscated commands using base64, hex encoding, and Unicode confusables
Hostname and transport layer analysis for suspicious network activity
If you need cloud/self-hosted management for an enterprise deployment, contact us.