Clampd
Runtime gateway that inspects every AI agent tool call against policy before execution to block injection, escalation and exfiltration
The product is a runtime gateway that sits between AI agents and tools to inspect every database query, API call, file access and shell command against policy before execution, blocking prompt injection, privilege escalation, data exfiltration and tool-chain attacks. It is for developers and IT/security teams in businesses that deploy AI agents with broad tool access. It is delivered as a self-hosted proxy with Python and TypeScript SDKs that evaluates 287 detection rules in microseconds and enforces scoped tokens and kill switches.
Key features
- Tool call firewall proxy
- 287 detection rules
- Prompt and response scanning
- SQL injection detection
- Command injection detection
- SSRF detection
- PII pattern detection
- Tool descriptor hash rug-pull detection
- Behavioural anomaly scoring per agent
- Cross-agent correlation
- Ed25519 scoped tokens
- Cedar policies with hot-reload
- AP2 and x402 payment validation
- Emergency kill cascade under 25ms
- Per-call audit logging
- SDK adapters for OpenAI Anthropic LangChain CrewAI MCP
- No social media activity within the last 30 days
GTM channels
- Blog
- API
- Docs
ICP
- Security teams
- Software developers
- IT teams