Skip to content
clawproof
$ ls ./navigation/
→ checks→ assessment→ open-source→ blog→ playbooksScore your agent →
$clawproof list --all

The 10 Clawproof Checks

Ten opinionated checks for AI agent reliability and governance, distilled from real production incidents. Each is a single failure mode you need to prevent before shipping.

Score your agent →Run them open source
Security3 checks
#01
Tool Permissions & Least Privilege
Your agent starts with read access. Within a month, it has keys to everything.
Read →#03
Prompt Injection & Data Exfil
Untrusted content in, secrets out. The attack surface nobody tests.
Read →#06
Secrets Management
API keys in prompts, tokens in logs. Zero secret sprawl or bust.
Read →
Governance2 checks
#04
Human-in-the-Loop & Escalation
When should an agent stop and ask? Define the boundaries before production.
Read →#08
Data Boundaries & RAG Governance
Your agent can read everything it retrieves. Can it read everything it should?
Read →
Operations2 checks
#05
Rollback & Kill Switches
Can you stop your agent in 30 seconds? If not, you're not production-ready.
Read →#09
Cost Controls & Rate Limiting
A runaway loop at 3AM. No spending cap. Monday morning surprise.
Read →
Quality3 checks
#02
Logging & Audit Trails
When something breaks, can you trace what happened, why, and who approved it?
Read →#07
Evaluation & Regression Testing
You upgrade the model. Agent behavior changes. No tests caught it.
Read →#10
Multi-Agent Coordination
Two agents, one resource, zero coordination. Race conditions aren't just for code.
Read →