A real AI architecture, not a chatbot bolted onto the UI.
RiskReady Community Edition exposes domain MCP servers, an orchestration gateway, approval-aware mutation paths, and the tasking foundation for autonomous workflows.
Mental model
254 tools across 9 MCP servers
Each domain exposes focused tools instead of hiding everything behind one generic assistant endpoint.
6-agent AI Council
Complex cross-domain questions convene 6 specialist agents (Risk Analyst, Controls Auditor, Compliance Officer, Incident Commander, Evidence Auditor, CISO Strategist) for structured deliberation.
Human-approved mutations
Write actions are proposed into an approval queue before they touch the database. This applies equally to interactive chat, scheduled runs, and autonomous workflows.
MCP Proxy for Claude Desktop
Connect Claude Desktop directly to your GRC data — bring your own AI. Remote MCP proxy with API key auth, zero AI cost to you.
Agentic gateway
The gateway coordinates routing, scheduled runs, cross-domain workflows with approval gates, and council-style analysis. Workflows pause and resume automatically around human decisions.
$0.19 per council (Haiku)
Full security posture assessment with 6 agents, 32 tool calls, 120K tokens. $0.007 per single query. 96% token reduction via tool search.
MCP surface
The community build ships nine domain-specific MCP servers.
Controls
66 toolsControls, SoA, assessments, metrics, and scope management.
Risks
34 toolsRisk registers, scenarios, KRIs, RTS, and treatment plans.
Policies
23 toolsDocuments, versions, approvals, reviews, mappings, and exceptions.
Organisation
35 toolsProfiles, committees, processes, departments, and governance data.
ITSM
40 toolsCMDB, change management, software inventory, and capacity planning.
Incidents
19 toolsIncidents, timelines, lessons learned, and related records.
Evidence
16 toolsEvidence repository, requests, linking, and analysis.
Audits
14 toolsNonconformities, audit findings, and corrective actions.
Agent Ops
7 toolsAction status, task tracking, workflow support, and operational introspection.
Mutations pause for review instead of silently writing.
The approval queue is central to the community story. It keeps the AI useful without hiding state changes behind a black box. For the full technical detail, use the AI Assistant and MCP reference docs in the repo.
Risk register
