Developer Hub

Build With ComplianceHarbor

Connect your AI agent to 48 risk intelligence tools via MCP. Compose workflows that combine ComplianceHarbor with GitHub, Jira, ServiceNow, Okta, and more.

Quick Start Guide Browse Tools Playbook Library

How It Works

Your AI agent orchestrates tools across multiple MCP servers in a single workflow. ComplianceHarbor provides risk intelligence; third-party servers provide operational context.

Your AI Agent

Claude, Copilot, GPT

↔

Third-Party MCP

GitHub, Jira, ServiceNow, Okta...

ComplianceHarbor MCP

48 risk intelligence tools

→

Risk-Aware Automation

Gates, evidence, alerts

Quick Start in 3 Steps

Connect your AI agent to ComplianceHarbor in under 2 minutes.

Configure MCP Client

// Claude Desktop - claude_desktop_config.json
{
  "mcpServers": {
    "complianceharbor": {
      "url": "https://complianceharbor.ai/mcp",
      "headers": {
        "X-API-Key": "ch_your_api_key"
      }
    }
  }
}
            

Also works with GitHub Copilot, Cursor, and any MCP-compatible client.

Run Your First Assessment

// Ask your AI agent:
"Assess the risk of deploying a
PostgreSQL 16 upgrade to production
this Saturday at 2am EST"

// Agent calls: assess_change_risk
// Returns: score, conflicts, evidence
            

The agent automatically formats the request and interprets results.

Combine with GitHub MCP

// Multi-MCP workflow:
// 1. GitHub MCP: get_pull_request
// 2. CH MCP: assess_change_risk
// 3. CH MCP: evaluate_rollback_trigger
// 4. CH MCP: generate_evidence_receipt
// 5. GitHub MCP: create_status_check

"For PR #142, assess deployment risk
and block if score exceeds 85"
            

See the DeployGuard playbook for the full workflow.

MCP Tool Reference

48 tools across 6 products. Filter by product to find the right tool for your workflow.

Loading tools from API...

Agent Playbook Library

Pre-built workflows that AI agents execute across ComplianceHarbor and third-party MCP servers.

Loading playbooks from API...

Compose Your Own Workflow

ComplianceHarbor tools accept input from any MCP context. The agent context normalizer automatically maps third-party field names to ComplianceHarbor parameters.

Field Alias Map

Source Field	Maps To	Platform
`repository`	`system`	GitHub
`ref`	`version`	GitHub
`workflow`	`changeDescription`	GitHub
`short_description`	`changeDescription`	ServiceNow
`cmdb_ci`	`affectedCIs`	ServiceNow
`planned_start_date`	`scheduledTime`	ServiceNow
`summary`	`changeDescription`	Jira
`components`	`affectedCIs`	Jira
`due_date`	`scheduledTime`	Jira
`hostname`	`system`	CrowdStrike
`display_name`	`system`	Okta
`search_name`	`changeDescription`	Splunk
`incident_name`	`changeDescription`	Sentinel

MCPToolResponse Schema

{
  "success": true,
  "tool": "assess_change_risk",
  "timestamp": "2026-03-10T...",
  "data": { /* tool-specific */ },
  "summary": "Risk score: 42/100",
  "riskScore": 42,
  "riskLevel": "medium",
  "conflicts": [...],
  "recommendations": [...],
  "sources": ["CISA KEV", ...],
  "agentHints": {
    "suggestedNextTools": [...],
    "requiredAction": false,
    "actionType": "review"
  }
}
            

Every tool returns a standardized envelope. Use agentHints.suggestedNextTools to chain tools into workflows.

Compatible MCP Servers

ComplianceHarbor workflows compose with these third-party MCP servers. Each has tested playbook integrations.

GitHub

PR gates, status checks

View playbooks →

GitLab

MR pipelines, deployments

Compatible

Atlassian (Jira)

Sprint risk, custom fields

View playbooks →

ServiceNow

CAB scoring, evidence sync

View playbooks →

Okta

Identity risk signals

View playbooks →

Slack

Compliance alerts

View playbooks →

Splunk

Threat enrichment

View playbooks →

Microsoft Sentinel

Incident context

View playbooks →

Ready to Build?

Get your API key, connect your AI agent, and start building risk-aware automation in minutes.

Get API Key API Documentation Browse Playbooks