Skip to main content
Glama

agentiam-mcp

Stdio MCP server wrapping all 18 AgentIAM x402 endpoints — agent identity, safety, and intelligence for autonomous agents on Base Mainnet.

What this is

agentiam-mcp exposes the AgentIAM HTTP catalog as a Model Context Protocol (MCP) server. Any MCP-capable client (Claude Desktop, Cursor, VS Code, LangChain, etc.) can call the 18 endpoints as tools.

Endpoints:

  • Safety pillars — noleak, memguard, riskoracle, secureexec, flowcore

  • Audit / policy — audit, validate, risk-check

  • Intelligence — research, intelligence-report

  • DELPHI knowledge graph — delphi, delphi.entity, delphi.query, delphi.timeline, delphi.contradictions

  • Signals — latest-signals, signal-query, publish-signal

Pricing: $0.001–$0.05 USDC per call, settled on Base via x402.

Related MCP server: obolpay-x402-mcp

Status

v0.1.0 — discovery mode. All 18 tools are registered. Tool calls return the AgentIAM endpoint's decoded x402 challenge (resource URL, accepts array, bazaar extension, schema), so MCP clients can browse the catalog without holding USDC. Wallet-signed payment flow (per-call USDC settlement) ships in v0.2.

Install

git clone https://github.com/achilliesbot/agentiam-mcp.git
cd agentiam-mcp
npm install

Run (standalone)

node src/index.mjs
# stdio MCP server — speaks JSON-RPC over stdin/stdout

Configure in Claude Desktop / Cursor / VS Code

Add to your MCP config (claude_desktop_config.json on macOS at ~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "agentiam": {
      "command": "node",
      "args": ["/absolute/path/to/agentiam-mcp/src/index.mjs"]
    }
  }
}

Restart the client. The 18 AgentIAM tools (noleak, memguard, riskoracle, secureexec, flowcore, audit, validate, risk_check, research, intelligence_report, delphi, delphi_entity, delphi_query, delphi_timeline, delphi_contradictions, latest_signals, signal_query, publish_signal) appear in the tool picker.

Run via Docker

docker build -t agentiam-mcp .
docker run -i agentiam-mcp

Smoke test

( echo '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"smoke","version":"0.1"}}}'
  sleep 0.3
  echo '{"jsonrpc":"2.0","method":"notifications/initialized"}'
  sleep 0.3
  echo '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'
  sleep 1
) | node src/index.mjs

You should see a JSON-RPC response listing 18 tools.

Discovery references

License

MIT.

Available Tools

18 tools
auditC

OWASP-class code safety audit. ($0.03 USDC per call on Base, via x402.)

Inputs: input.

delphiC

Real-time intel signals (crypto, AI, DeFi, macro). ($0.01 USDC per call on Base, via x402.)

Inputs: input.

delphi_contradictionsC

Conflicting intelligence detection. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

delphi_entityB

Knowledge graph entity lookup + relationships. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

delphi_queryC

Graph query by predicate/subject/object. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

delphi_timelineB

Chronological fact history for any entity. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

flowcoreA

Full orchestration pipeline — chains NoLeak + MemGuard + RiskOracle + SecureExec. ($0.02 USDC per call on Base, via x402.)

Inputs: input.

intelligence_reportC

Deep DELPHI synthesis — cross-referenced signals. ($0.05 USDC per call on Base, via x402.)

Inputs: input.

latest_signalsC

Cheapest entry — latest signals across all categories. ($0.001 USDC per call on Base, via x402.)

Inputs: input.

memguardB

Verify memory/state — detect drift, corruption, unauthorized modifications. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

noleakB

Verify execution integrity — detect leaks, injection, unauthorized data flow. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

publish_signalB

Publish a signal — earn 70% of downstream query fees. ($0.005 USDC per call on Base, via x402.)

Inputs: input.

researchC

Structured intel brief — web research + analysis. ($0.05 USDC per call on Base, via x402.)

Inputs: input.

risk_checkC

Lightweight risk score by action type/value/leverage. ($0.005 USDC per call on Base, via x402.)

Inputs: input.

riskoracleC

Pre-action risk scoring — multi-factor analysis before executing an action. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

secureexecB

Sandboxed tool execution with proof hash. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

signal_queryA

Query signals by type/severity/keyword/time. ($0.002 USDC per call on Base, via x402.)

Inputs: input.

validateA

Deterministic policy validation — risk + compliance + proof hash. ($0.01 USDC per call on Base, via x402.)

Inputs: input.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3/5.0

    Scored across 18 tools

    Disambiguation4/5

    Most tools have distinct purposes, especially the delphi_* cluster with specific query types. However, 'delphi' and 'latest_signals' both deal with real-time signals and could cause confusion, though descriptions differentiate them. Similarly, 'risk_check' and 'riskoracle' overlap in risk assessment but with different scopes. Overall, agents can likely select the correct tool with careful reading.

    Naming Consistency3/5

    Naming uses a mix of conventions: single words ('audit', 'research'), concatenated words ('flowcore', 'riskoracle'), and underscores ('delphi_entity', 'risk_check'). The delphi_* tools follow a consistent pattern, but the rest do not follow a uniform verb_noun or other clear pattern. This mixed style reduces predictability.

    Tool Count4/5

    18 tools is on the higher side but reasonable for the broad domain covering security audits, intelligence signals, risk analysis, and execution verification. Each tool serves a specific function, and the count does not feel excessive given the scope. However, a slight reduction could improve coherence.

    Completeness4/5

    The tool set covers most core workflows: signal querying, publishing, risk assessment, verification, and orchestration. Minor gaps exist, such as lacking signal update/delete tools and possibly entity management for the knowledge graph. Overall, the surface is largely complete for security and intelligence tasks.

    Related MCP Connectors

    Related MCP Servers