MCP 服务器

DCL Trust Oracle

DCL Trust Oracle is a deterministic AI audit layer, natively integrated with MCP, that evaluates LLM and agent outputs against configurable policies before and after action. Every verdict is written to a tamper-evident, hash-chained audit log storing only cryptographic metadata — never raw content — enabling privacy-first forensic review.

README

DCL Trust Oracle — x402 MCP Server

Deterministic AI audit layer with cryptographic micropayments via x402 protocol.

What It Does

DCL Trust Oracle provides deterministic policy evaluation for LLM outputs with a tamper-evident audit chain. The system stores only cryptographic hashes and decision metadata — never raw content — enabling verifiable, post-action forensic analysis across distributed AI agents.

Available two ways, both metered per call via x402:

REST API (webhook_server.py) — direct HTTP integration, x402-gated with fastapi-x402.
MCP Server (mcp_server.py) — native Model Context Protocol integration for AI agents, hosted on Smithery, x402-gated with paymcp.

Both servers share the same evaluation logic and tamper-evident chain (dcl_core.py), and are priced identically.

Quick Start

REST API

pip install -r requirements.txt
python webhook_server.py

Server runs on http://localhost:8080

MCP Server

pip install -r requirements.txt
python mcp_server.py

Server runs on http://localhost:8081 (streamable-http transport)

Tools & Endpoints

Pre-Action Evaluation

REST Endpoint	MCP Tool	Price	Description
`POST /evaluate/fast`	`dcl_evaluate_fast`	$0.01	Fast policy check for low-risk outputs. Returns tamper-evident `tx_hash`.
`POST /evaluate/strict`	`dcl_evaluate_strict`	$0.05	Deep analysis for high-stakes outputs with higher confidence thresholds.
`POST /evaluate/jailbreak`	`dcl_evaluate_jailbreak`	$0.02	Instruction adherence check — detects prompt injection patterns and role-hijacking attempts.
`POST /evaluate/safety`	`dcl_evaluate_safety`	$0.01	Baseline screening for known harmful text patterns. Optimized for high throughput.
`POST /evaluate/quality`	`dcl_evaluate_quality`	$0.03	Content quality & drift check — evaluates format adherence and contextual drift.
`POST /evaluate/batch`	`dcl_evaluate_batch`	$0.10	Bulk processing — up to 20 items per transaction. Cost-effective for multi-turn history.

Session Management

REST Endpoint	MCP Tool	Price	Description
`POST /pipeline/start`	`dcl_pipeline_start`	$0.05	Initializes a long-running audit session for continuous drift tracking. Returns `pipeline_id`.

Post-Action Forensics

REST Endpoint	MCP Tool	Price	Description
`GET /audit/{tx_hash}`	`dcl_audit_decode`	$0.10	Basic post-action audit — returns verdict, confidence, agent_id, reason by `tx_hash`.
`GET /audit/{tx_hash}/deep`	`dcl_audit_decode_deep`	$0.50	Deep forensic audit — includes drift context, tamper-evidence indices, environmental metadata.

Utility (free, REST only)

Endpoint	Description
`GET /health`	Service status and chain length
`GET /policies`	List of built-in policy names
`GET /chain/status`	Chain integrity, drift mode, drift score
`GET /chain/export`	Full chain export with integrity verification

Example Response

{
  "verdict": "COMMIT",
  "confidence": 0.95,
  "reason": "All policy checks passed",
  "tx_hash": "0x7a8f3b2c...",
  "chain_index": 42,
  "input_hash": "0x9d4e1f...",
  "policy_version": "1.0.0",
  "timestamp": 1721635200.123,
  "pipeline_id": "abc123",
  "drift_mode": "NORMAL",
  "drift_score": 0.15
}

x402 Integration

All paid endpoints and tools require payment via the x402 protocol before returning a result — no free tier, no bypass. Both the REST API and MCP server settle to the same wallet.

Networks: Base, Avalanche, IoTeX
Asset: USDC
Facilitator (REST): https://x402.org/facilitator (via fastapi-x402)
Facilitator (MCP): paymcp in Mode.AUTO — automatic on-chain payment for x402-aware MCP clients, guided payment link (ELICITATION/RESUBMIT) for clients without a wallet. No path skips payment.

Metadata-Only Architecture

DCL Trust Oracle is designed around a hash-based audit trail:

Raw content never stored — only SHA-256 hashes and decision metadata
Tamper-evident cryptographic chain (SQLite-backed, WAL mode)
Chain survives server restarts — audit records persist indefinitely
Verifiable integrity via GET /chain/status
Full export via GET /chain/export
Shared chain across REST and MCP — a tx_hash returned by one interface can be audited through the other

Each audit record contains: tx_hash, prev_hash, verdict, input_hash, policy_hash, agent_id, reason, confidence, task_type, timestamp, and drift_context.

Built-in Policies

Policy	Min Confidence	Purpose
`default`	0.70	General-purpose evaluation
`anti_jailbreak`	0.80	Detects prompt injection, role-hijacking, DAN-style attacks
`safety`	0.75	Baseline harmful-pattern screening
`content_quality`	0.85	Quality assurance, format adherence, drift detection