Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@sigmodx-mcp Log invoice approval for PO-12345 for $5000"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server for Sigmodx — audit infrastructure for AI agents making consequential decisions.
Exposes Sigmodx as MCP tools any compatible agent framework can call. Works with Claude, Cursor, LangChain, AutoGen, and any MCP client.
pip install sigmodx-mcpRelated MCP server: cronozen-proof
SIGMODX_API_KEY=your-key \
SIGMODX_AGENT_ID=your-agent-uuid \
sigmodx-mcpAdd to your MCP config:
{
"mcpServers": {
"sigmodx": {
"command": "sigmodx-mcp",
"env": {
"SIGMODX_API_KEY": "your-api-key",
"SIGMODX_AGENT_ID": "your-agent-uuid"
}
}
}
}Tool | Description |
| Log invoice approval/rejection |
| Log GL entry review decision |
| Log anomaly flag/clear/escalate |
| Verify a verification string |
| Get agent ALLOW/LIMIT/BLOCK state |
| Hash input payload (no auth needed) |
Get the current reliability state (ALLOW/LIMIT/BLOCK) for an agent in a specific scenario.
Inputs: scenario (required).
Hash an input payload using Sigmodx's deterministic SHA-256 method. The payload itself is never sent to Sigmodx.
Inputs: payload (required).
Log an anomaly detection decision to Sigmodx. Use when an AI agent flags, clears, or escalates a financial transaction anomaly. Critical severity items are automatically escalated for immediate review.
Inputs: decision_type (required), inputs (required), rationale (required), anomaly_subtype, +5 more.
Log a GL entry review decision to Sigmodx. Use when an AI agent approves, flags, or blocks a journal entry. Segregation of duties violations are auto-blocked.
Inputs: decision_type (required), inputs (required), rationale (required), flag_subtype, +4 more.
Log an invoice approval decision to Sigmodx for cryptographic attestation. Use when an AI agent approves, rejects, or escalates an invoice. The input payload is hashed client-side — invoice data never leaves your environment.
Inputs: decision_type (required), inputs (required), rationale (required), invoice_amount, +2 more.
Verify a Sigmodx attestation using a verification string. Returns the attestation record and confirms the cryptographic hash is intact. No authentication required.
Inputs: verification_string (required).
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 6 tools
Each tool targets a distinct operation: reliability state, input hashing, three decision logging types (anomaly, GL, invoice), and attestation verification. No functional overlap exists.
All tools follow a consistent 'sigmodx_verb_noun' pattern (e.g., get_reliability, hash_inputs, log_anomaly_decision). Naming is uniform and predictable.
With 6 tools, the set is well-scoped for a specialized MCP server focusing on Sigmodx operations. Each tool serves a clear purpose without being too few or too many.
The tool surface covers core operations: reliability checking, input hashing, decision logging for three key domains, and attestation verification. Minor gaps like a general decision retrieval tool are absent but not critical for the stated purpose.
Bitcoin-anchored, tamper-evident audit log for AI agents — record, disclose and verify actions.
AI/LLM agent output audit MCP: policy eval, tamper-evident chain, AI safety, x402 USDC on Base.
AI agent security, MCP 2026 schema & graph audits, <250ms guardrails, and compliance dossiers
Policy-aware AI execution, verification, cost, and trajectory evidence for agents.