GAIA-Ops
General Agentic Integration Architecture
Multi-agent DevOps system that classifies every operation by risk, routes work to specialist agents, and blocks irreversible commands automatically.
Overview
GAIA-Ops is a multi-agent orchestration system for DevOps automation. It provides security-first command classification, specialized AI agents, and plugin-based distribution. Currently integrates with Claude Code.
Features
- Multi-cloud support - GCP, AWS, Azure
- 6 agents - terraform-architect, gitops-operator, cloud-troubleshooter, devops-developer, speckit-planner, gaia-system (meta-agent)
- Contracts as SSOT - Cloud-agnostic base contracts with per-cloud extensions (GCP, AWS)
- Dynamic identity - Orchestrator identity injected by UserPromptSubmit hook; skills loaded on-demand
- Dual-barrier security - Settings deny rules (Claude Code native) + hook-level blocking (inalterable via symlink)
- Indirect execution detection - Catches
bash -c,eval,python -cwrappers that bypass regex patterns - Approval gates for T3 operations via native
askdialog - Git commit validation with Conventional Commits
- 21 skills - Injected procedural knowledge modules for agents
- Plugin + npm - Distributable as Claude Code native plugin or npm package
- Enterprise ready - Managed settings template for organization-wide deployment
Installation
Via Claude Code Plugin (recommended)
# Add the marketplace
/plugin marketplace add metraton/gaia-ops
# Install the full system (includes security)
/plugin install gaia-ops
# Or install security only
/plugin install gaia-security # Security hooks only
Via npm (advanced setup)
npm install @jaguilar87/gaia-ops
npx gaia-scan
Quick Start (npm)
# Run directly with npx
npx gaia-scan
# Or install globally
npm install -g @jaguilar87/gaia-ops
gaia-scan
This will:
- Auto-detect your project structure (GitOps, Terraform, AppServices)
- Create
.claude/directory with symlinks to this package - Generate
project-context.json - Create
settings.jsonwith hooks only (no permissions in settings.json) - Merge deny rules + allow permissions into
settings.local.json(preserves existing user config)
No CLAUDE.md is generated -- orchestrator identity is injected dynamically by the UserPromptSubmit hook.
Settings Architecture
Gaia-Ops separates hooks from permissions:
| File | Content | Strategy |
|---|---|---|
settings.json | Hooks only (9 hook types) | Overwritten from template on each update |
settings.local.json | Permissions (allow + deny rules) | Union merge — never removes user config |
This ensures your personal customizations (MCP servers, extra permissions) survive updates.
Manual Installation
npm install @jaguilar87/gaia-ops
Then create symlinks:
mkdir -p .claude && cd .claude
ln -s ../node_modules/@jaguilar87/gaia-ops/agents agents
ln -s ../node_modules/@jaguilar87/gaia-ops/tools tools
ln -s ../node_modules/@jaguilar87/gaia-ops/hooks hooks
ln -s ../node_modules/@jaguilar87/gaia-ops/commands commands
ln -s ../node_modules/@jaguilar87/gaia-ops/config config
ln -s ../node_modules/@jaguilar87/gaia-ops/templates templates
ln -s ../node_modules/@jaguilar87/gaia-ops/skills skills
ln -s ../node_modules/@jaguilar87/gaia-ops/speckit speckit
Usage
Once installed, the agent system is ready:
claude
The orchestrator identity is injected dynamically by the UserPromptSubmit hook. Skills are loaded on-demand.
Skills and injection diagnosis:
npx gaia-skills-diagnose
# or with test probe:
npx gaia-skills-diagnose --run-tests
Security
Gaia-Ops enforces a 6-layer security pipeline:
| Layer | Mechanism | Bypassable? |
|---|---|---|
| Indirect execution detection | bash -c, eval, python -c wrappers | No (hook-level) |
| Blocked commands (regex) | 85+ regex patterns | No (symlink to npm package) |
| Blocked commands (semantic) | 70+ ordered-token rules | No (symlink to npm package) |
| Cloud pipe validator | Credential piping detection | No (hook-level) |
| Mutative verb detection | ask dialog for state-changing ops | User approves via native dialog |
| Settings deny rules | 147 deny rules in settings.local.json | Self-healing (restored each session) |
Enterprise Deployment
For organization-wide enforcement, deploy templates/managed-settings.template.json as a managed settings policy via Claude.ai Admin Console. Managed settings have the highest precedence and cannot be overridden.