Summary
This is a full governance layer for AI agents built around the MAI decision framework (Mandatory, Advisory, Informational). You get 29 tools split across decision gates, compliance mapping (NIST AI RMF, EU AI Act, ISO 42001), hash-chained audit ledgers, and knowledge pack management with trust levels and TTLs. The engine runs embedded with optional Postgres for persistence, enforcing human-in-the-loop approvals at mandatory gates and generating forensic audit trails you can cryptographically verify. Reach for this if you're deploying agents in regulated environments (FedRAMP, CMMC, SOC 2) and need enforceable controls, not just logging. Also includes infrastructure remediation packs and economic ROI tracking for governance interventions.
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →
GIA Governance Intelligence Automation
Enterprise AI governance through the Model Context Protocol.
GIA is a production governance engine that gives AI agents enforceable decision controls, compliance scoring, immutable audit chains, and human-in-the-loop gates. Built for organizations operating under NIST, FedRAMP, CMMC, EU AI Act, and SOC 2 requirements.
29 MCP tools. One integration point. Integrates with any MCP-compatible agent, framework, or client.
Quick Start
The governance engine runs fully embedded — no API key, no external dependency.
npx gia-mcp-server
Or install globally:
npm install -g gia-mcp-server
gia-mcp-server
All 29 tools are available immediately. By default the engine runs in-memory. Add a Postgres connection string for persistent audit trails:
DATABASE_URL=postgresql://... npx gia-mcp-server
MCP Client Configuration
Add to your MCP client config (stdio transport):
{
"mcpServers": {
"gia-governance": {
"command": "npx",
"args": ["-y", "gia-mcp-server"]
}
}
}
Streamable HTTP (Remote)
A hosted endpoint is available for frameworks that prefer HTTP transport:
https://gia.aceadvising.com/mcp
Requires GIA_API_KEY — contact aceadvising.com for access.
Smithery
npx -y @smithery/cli install gia-mcp-server
Tools
Decision Controls (MAI Framework)
| Tool | Description |
|---|---|
classify_decision | Classify agent decisions as Mandatory, Advisory, or Informational |
approve_gate | Human-in-the-loop approval for Mandatory gates |
evaluate_threshold | Compute escalation health (Storey Threshold) |
score_governance | Weighted governance scoring (Integrity, Accuracy, Compliance) |
Compliance & Audit
| Tool | Description |
|---|---|
audit_pipeline | Query the hash-chained forensic audit ledger |
verify_ledger | Verify SHA-256 chain integrity from genesis |
map_compliance | Map controls to NIST AI RMF, EU AI Act, ISO 42001, NIST 800-53 |
assess_risk_tier | EU AI Act risk tier classification |
generate_report | Governance status reports (summary, detailed, executive) |
Knowledge Packs
| Tool | Description |
|---|---|
seal_memory_pack | Create immutable, TTL-bound knowledge artifacts |
load_memory_pack | Load packs with trust level and role enforcement |
transfer_memory_pack | Governed knowledge transfer between agents |
compose_memory_packs | Merge packs with risk contamination rules |
distill_memory_pack | Extract governance patterns from usage history |
promote_memory_pack | Promote packs to higher trust levels after review |
Security & Operations
| Tool | Description |
|---|---|
monitor_agents | Agent health, repair history, failure counts |
srt_run_watchdog | Infrastructure health probes (API, disk, memory, TLS, DB, DNS) |
srt_diagnose | Incident diagnosis with playbook matching |
srt_approve_repair | Human-approved repair execution |
srt_generate_postmortem | Structured incident postmortems with TTD/TTR metrics |
Infrastructure Remediation
| Tool | Description |
|---|---|
gia_scan_environment | Scout swarm for environment detection |
gia_list_packs | List remediation, patrol, hardening, and audit packs |
gia_dry_run_pack | Preview pack execution with blast radius analysis |
gia_apply_pack | Execute remediation with mandatory human approval |
gia_run_patrol | Read-only posture checks and compliance audits |
Impact & Value
| Tool | Description |
|---|---|
record_value_metric | Track time saved, risks blocked, autonomy levels |
record_governance_event | Log gates, drift prevention, violations blocked |
generate_impact_report | Economic + governance ROI reporting |
system_status | Engine health, uptime, configuration |
Architecture
GIA enforces governance through three layers:
- Decision Controls — MAI classification gates side effects and high-impact actions
- Step Hooks — Workflow progression control at each pipeline stage
- Kernel Hooks — Resource control at the LLM boundary, including sub-agents
Every governance action is recorded in a SHA-256 hash-chained audit ledger that can be independently verified.
Compliance Coverage
- NIST AI RMF — Risk management framework mapping
- EU AI Act — Risk tier assessment and control mapping
- ISO 42001 — AI management system alignment
- NIST 800-53 — Federal security control mapping
- CMMC 2.0 — DoD cybersecurity maturity
- FedRAMP — Federal cloud authorization
- SOC 2 — Service organization controls
About
Built by Advanced Consulting Experts (ACE) — a Service-Disabled Veteran-Owned Small Business (SDVOSB).
GIA was designed by William J. Storey III, a 17-year Information System Security Officer with experience across DoD contracts and U.S. Army Ranger Battalion operations. The same discipline applied to securing classified systems now governs AI agent workforces.
License
MIT
Featured
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →Categories
Registryactive
Packagegia-mcp-server
TransportSTDIO, HTTP
AuthRequired
UpdatedMar 12, 2026
