Airis Agent
Autonomous AI workflow orchestrator for Claude Code
Airis Agent is the intelligence layer of the Airis Suite, providing autonomous workflow orchestration with confidence gating, deep research, repository indexing, and self-review capabilities for AI-enhanced development.
🌟 Part of the AIRIS Ecosystem
Airis Agent is the intelligence layer of the AIRIS Suite - making all AI coding assistants smarter across Claude Code, Cursor, Windsurf, Codex, and Gemini CLI.
The AIRIS Suite
| Component | Purpose | For Who |
|---|
| airis-agent (this repo) | 🧠 Intelligence layer for all editors (confidence checks, deep research, self-review) | All developers using Claude Code, Cursor, Windsurf, Codex, Gemini CLI |
| airis-mcp-gateway | 🚪 Unified MCP proxy with 90% token reduction via lazy loading | Claude Code users who want faster startup |
| mindbase | 💾 Local cross-session memory with semantic search | Developers who want persistent conversation history |
| airis-workspace | 🏗️ Docker-first monorepo manager | Teams building monorepos |
| airiscode | 🖥️ Terminal-first autonomous coding agent | CLI-first developers |
MCP Servers (Included via Gateway)
- airis-mcp-supabase-selfhost - Self-hosted Supabase MCP with RLS support
- mindbase - Memory search & storage tools (
mindbase_search, mindbase_store)
Quick Install: Complete AIRIS Suite
# Option 1: Install airis-agent plugin (recommended for Claude Code users)
/plugin marketplace add agiletec-inc/airis-agent
/plugin install airis-agent
# Option 2: Clone all AIRIS repositories at once
uv run airis-agent install-suite --profile core
# Option 3: Install individual components as needed
# See each repository for standalone installation
What you get with the full suite:
- ✅ Confidence-gated workflows (prevents wrong-direction coding)
- ✅ Deep research with evidence synthesis
- ✅ 94% token reduction via repository indexing
- ✅ Cross-session memory across all editors
- ✅ Self-review and post-implementation validation
What is Airis Agent?
Airis Agent is an autonomous workflow orchestrator that enhances Claude Code with:
- Pre-implementation Confidence Checks - Validates readiness before coding (≥90% confidence required)
- Deep Research - Parallel web search with evidence-based synthesis using Tavily/Context7
- Repository Indexing - 94% token reduction through intelligent codebase analysis
- Self-Review - Post-implementation validation with evidence requirements
Philosophy: Prevent wrong-direction work through confidence gating, not post-hoc fixes.
Installation
Prerequisites
Required:
- Python 3.10+ with
uvx (auto-installed by Claude Code for MCP servers)
- Both
airis-agent-mcp and airis-mcp-gateway use uvx for lightweight deployment
- No Docker required for basic usage
Optional but Recommended:
airis-workspace - Monorepo management tool (enhances airis-agent capabilities)airiscode - Terminal-first autonomous coder
See Install the complete Airis Suite below for one-command installation.
From GitHub Marketplace (Recommended)
Step 1: Add the Agiletec marketplace to Claude Code
/plugin marketplace add agiletec-inc/airis-agent
Step 2: Install the plugin
/plugin install airis-agent
Step 3: Verify installation
/plugin list
You should see: airis-agent@agiletec-inc (enabled)
Step 4: Test the plugin
# All workflows are handled by the airis_agent MCP tool with automatic intent detection
# The tool detects intent from your task description:
# - "research JWT best practices" → Deep Research Protocol
# - "index repository" → Repository Index Protocol
# - "optimize docs" → Document Optimization Protocol
# - Other tasks → Task Protocol (with Confidence Check)
MCP-First Architecture: No slash commands needed. The airis_agent MCP tool orchestrates all workflows through intelligent intent detection
Local Development