AI governance plugins for Claude Code. Safety classification, compliance automation, forensic analysis, and governed operations — powered by the GIA (Governed Intelligence Architecture) engine.
# Add the marketplace
/plugin marketplace add knowledgepa3/gia-marketplace
# Install a free plugin
/plugin install gia-safety-guard@gia-marketplace
# Install a pro plugin (requires API key)
/plugin install gia-compliance-ops@gia-marketplace| Plugin | What It Does |
|---|---|
| gia-safety-guard | Classify every AI decision as Mandatory, Advisory, or Informational. Prevent agent drift before it starts. |
| ace-forensic-analyst | VA disability claims forensic analysis. Shedden elements, nexus language, evidence chain verification. |
| Plugin | What It Does |
|---|---|
| gia-compliance-ops | Full AI compliance suite. /classify, /threshold, /score, /audit, /comply, /report — with live MCP server. |
| ace-claims-pipeline | Governed forensic claims pipeline. /ingest, /analyze, /red-team, /gate, /export, /status — with audit trails. |
| stack-governed-ops | Full-stack governance. 23 MCP tools, governed engineer agent, enforcement hooks, memory packs. The complete GIA experience. |
Skills and agents that run locally in Claude Code. No server connection needed. You get the governance mindset — MAI classification, forensic analysis frameworks, safety awareness.
Commands and tools that connect to the live GIA MCP server at gia.aceadvising.com. You get real enforcement — audit trails, compliance mapping, mandatory gates, memory packs, and 23 governance tools.
Free Plugin Pro Plugin
┌─────────────┐ ┌─────────────────────────┐
│ SKILL.md │ │ SKILL.md + Commands │
│ Agent.md │ │ .mcp.json ──────────┐ │
│ │ │ hooks.json │ │
│ (local) │ │ │ │
└─────────────┘ └───────────────────────│──┘
│
▼
┌───────────────────┐
│ GIA MCP Server │
│ gia.aceadvising │
│ .com/mcp │
│ │
│ 23 governance │
│ tools, audit │
│ ledger, memory │
│ packs, gates │
└───────────────────┘
Pro plugins require a GIA_API_KEY environment variable.
- Visit gia.aceadvising.com
- Create an account
- Subscribe to a plan
- Copy your API key
- Set it in your environment:
# Add to your shell profile (.bashrc, .zshrc, etc.)
export GIA_API_KEY=gia_your_key_hereSee docs/API-KEY-SETUP.md for detailed setup instructions.
| Plan | Rate Limit | Tool Calls/Day | Sessions | Price |
|---|---|---|---|---|
| Starter | 30 req/min | 1,000/day | 3 concurrent | Free trial |
| Professional | 120 req/min | 10,000/day | 10 concurrent | $49/mo |
| Enterprise | 600 req/min | Unlimited | Unlimited | Contact us |
GIA maps to established regulatory frameworks:
| Framework | What It Covers |
|---|---|
| NIST 800-53 Rev 5 | Federal security controls (AC, AU, SI families) |
| NIST AI RMF 1.0 | AI risk management (GOVERN, MAP, MEASURE) |
| EU AI Act | European AI regulation, risk tier assessment |
| ISO 42001 | AI management systems |
Every AI decision is classified into one of three categories:
| Level | Name | Behavior |
|---|---|---|
| M | Mandatory | Hard-coded rules. Cannot be bypassed. Block or require human approval. |
| A | Advisory | Guidelines. Follow by default. Override requires justification (logged). |
| I | Informational | Awareness. No enforcement. Logging and monitoring only. |
This classification is the foundation of all GIA governance. Whether you're using the free Safety Guard or the full GovOps Stack, every action is classified before execution.
ACE Advising — Governed Intelligence Architecture (GIA)
- Platform: gia.aceadvising.com
- GitHub: knowledgepa3
