TrustMesh LangChain

LangChain tools for TrustMesh — AI agent trust & reputation scoring.

Install

npm install trustmesh-langchain

Quick Start

import { getTrustMeshTools } from 'trustmesh-langchain';

const tools = getTrustMeshTools();
// Returns: [TrustScoreTool, TrustRateTool, TrustCheckTool, TrustGraphTool, TrustDisputeTool]

Or use individual tools:

import { TrustScoreTool, TrustCheckTool, TrustRateTool } from 'trustmesh-langchain';

const scoreTool = new TrustScoreTool();
const checkTool = new TrustCheckTool();
const rateTool = new TrustRateTool();

With LangChain Agent

import { ChatOpenAI } from '@langchain/openai';
import { AgentExecutor, createOpenAIFunctionsAgent } from 'langchain/agents';
import { ChatPromptTemplate } from '@langchain/core/prompts';
import { getTrustMeshTools } from 'trustmesh-langchain';

const tools = getTrustMeshTools();
const llm = new ChatOpenAI({ modelName: 'gpt-4o' });

const prompt = ChatPromptTemplate.fromMessages([
  ['system', 'You are a trust-aware AI agent. Check trust scores before engaging with other agents.'],
  ['human', '{input}'],
  ['placeholder', '{agent_scratchpad}'],
]);

const agent = await createOpenAIFunctionsAgent({ llm, tools, prompt });
const executor = new AgentExecutor({ agent, tools });

// The agent will automatically check trust before transactions
const result = await executor.invoke({ input: 'Should I delegate this task to agent-456?' });

Tools

TrustScoreTool (trustmesh_score)

Get trust score for an agent. Returns overall score (0-1), confidence, and sample size.

TrustRateTool (trustmesh_rate)

Rate an interaction with another agent. Score 0-1 with optional claim.

TrustCheckTool (trustmesh_check)

Quick trust check — returns proceed, caution, or avoid.

TrustGraphTool (trustmesh_graph)

Get trust graph around an agent — who trusts them, who they trust.

TrustDisputeTool (trustmesh_dispute)

Dispute a fraudulent or unfair rating.

Environment Variables

| Variable | Default | Description | |----------|---------|-------------| | TRUSTMESH_NODE_ID | langchain-<random> | Your agent's identity | | TRUSTMESH_REMOTE_URL | local | Remote TrustMesh API URL |

Links

• npm: https://www.npmjs.com/package/trustmesh-langchain
• GitHub: https://github.com/geerdwedda-create/trustmesh
• Core SDK: https://www.npmjs.com/package/trustmesh-ai
• MCP Server: https://www.npmjs.com/package/trustmesh-mcp
• Docs: https://geerdwedda-create.github.io/trustmesh/

License

MIT