Building the cryptographic verification layer for the AI agent economy.
Don't trust the agent. Trust the proof.
A deterministic cryptographic protocol for sealing, verifying, and auditing AI agent decisions.
Every action an agent takes can be committed into a tamper-evident hash chain — privacy-first (raw content is never stored, only hashes).
Production implementation of Leibniz Layer™ as a live MCP server.
Metered via x402 (USDC on Base). No account. Pay per call.
→ MCP: https://mcp.fronesislabs.com/mcp
→ Smithery · Glama · 402 Index
Client-side verification for the DCL protocol — no server round-trip required. Recomputes the hash locally and tells you if a chain record was tampered with.
→ TypeScript: npm install @fronesis-labs/dcl-sdk — npmjs.com/package/@fronesis-labs/dcl-sdk · github.com/Fronesis-Labs/dcl-sdk
→ Python: pip install dcl-core — pypi.org/project/dcl-core · github.com/Fronesis-Labs/dcl-core
Instruction-only + live-mode skills for AI agents:
jailbreak defense, PII/secret redaction, policy enforcement, hallucination checks, and a DeFi/crypto suite.
→ github.com/Fronesis-Labs/dcl-skills
→ ClawHub @daririnch
License: Apache 2.0 (this and related repos).
ClawHub listings are under the platform's required MIT-0.
```json { "mcpServers": { "dcl-trust-oracle": { "url": "https://mcp.fronesislabs.com/mcp" } } } ```
Payment is handled via x402 (USDC on Base). No API key or account required for x402-capable clients.
| 🌐 Website | fronesislabs.com |
| ⬡ Leibniz Layer™ | leibniz.fronesislabs.com |
| 📦 DCL Skills | github.com/Fronesis-Labs/dcl-skills |
| 🔌 MCP / Trust Oracle | github.com/Fronesis-Labs/dcl-webhook |
| 🛠️ Smithery | smithery.ai/servers/fronesislabs/dcl-trust-oracle |
| 🦙 Glama | glama.ai/mcp/servers/Fronesis-Labs/dcl-webhook |
| 📄 Whitepaper | leibniz.fronesislabs.com |
| 🐦 X | @keykeeper42 |
| 📧 Contact | partnership@fronesislabs.com |
Every decision, every action — deterministically sealed, tamper-evident, auditable.
© 2026 Fronesis Labs · Open source (Apache 2.0)