Claude-Cortex / site

# Cortex

> Multi-model development orchestration for Claude Code, Codex, and Gemini

Cortex is a context orchestration framework that coordinates AI coding agents across model providers, shipped as a Python CLI plus Textual TUI. It bundles curated agents, skills, rules, and hooks for Claude Code, and enforces quality gates — independent code review, test-coverage audits, lint checks — so that no agent grades its own homework.

## Core principle

The agent that writes the code never reviews it. When one model implements a feature, the review is routed to a different model family first, then to a different model in the same family, with same-instance review only as a last resort. Every review produces a structured artifact with severity levels and a pass/fail verdict.

## Three quality loops

- **Code change loop** — implement → independent review → remediate P0/P1 → re-review (max 3 cycles)
- **Test writing loop** — audit gaps → write tests → verify → re-audit (max 3 cycles)
- **Lint gate** — discover linter → auto-fix → check → remediate (max 2 cycles)

P0/P1 findings block the loop. P2/P3 are filed as backlog issues. Circuit breakers escalate to a human rather than looping infinitely.

## What ships

| Asset | Purpose |
|---|---|
| `agents/` | Specialized subagents (review, security, architecture, language pros) |
| `skills/` | Skill modules with progressive disclosure |
| `rules/` | Behavioral guardrails and coding conventions |
| `hooks/` | Claude Code lifecycle automation (skill suggestion, validation) |
| `claude_ctx_py/` | Python CLI and Textual TUI |

## Skill suggestion (not activation)

Skills surface as you work via a two-layer pipeline: fast keyword matching on every prompt (~50ms), with optional semantic matching (`fastembed`) for deeper recommendations. The TUI runs a background watch daemon for continuous suggestions. Cortex suggests; it does not auto-activate — the user or agent decides whether to invoke a suggested skill.

Agents, by contrast, stay active once installed.

## Install

```
# Python (any platform)
pipx install claude-cortex

# macOS (Homebrew)
brew tap NickCrew/cortex
brew install cortex
```

Then link bundled assets into `~/.claude/` and launch the TUI:

```
cortex install link
cortex tui
```

## CLI surface

```
cortex status                          active agents, rules, hooks, skills
cortex agent list|status               manage agent definitions
cortex skills list|info|recommend      discover and inspect skills
cortex rules list|activate             manage behavioral rules
cortex hooks list|install              install and validate hooks
cortex mcp list|diagnose               MCP server discovery and diagnostics
cortex ai recommend|watch              skill recommendations (one-shot or daemon)
cortex review                          run review workflows
cortex tui                             launch terminal UI
cortex docs                            browse bundled documentation
```

`cortex <command> --help` shows the full surface for any subcommand.

## Scope resolution

Cortex resolves its data folder in this order:

1. `CORTEX_SCOPE` (`project` / `global` / `plugin`)
2. `CLAUDE_PLUGIN_ROOT` (set by Claude Code for plugin commands)
3. `CORTEX_ROOT` (default: `~/.cortex`)

Active assets are tracked in `.active-*` state files (`.active-agents`, `.active-rules`, etc.) — not in CLAUDE.md comments.

## Links

- Repository: https://github.com/NickCrew/claude-cortex
- Documentation: https://cortex.atlascrew.dev
- Full reference for LLMs: https://cortex.atlascrew.dev/llms-full.txt
- PyPI: https://pypi.org/project/claude-cortex/
- Issues: https://github.com/NickCrew/claude-cortex/issues
- Changelog: https://github.com/NickCrew/claude-cortex/blob/main/CHANGELOG.md

## License

MIT