Skip to content

rubakas/agent-notes

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

65 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

agent-notes

AI agent configuration manager for Claude Code and OpenCode.

Configures a Lead agent (Opus) that orchestrates a team of 13 specialized subagents across three model tiers — so Opus plans, Sonnet executes, and Haiku explores.

Quick Start

git clone https://github.com/rubakas/agent-notes.git ~/agent-notes
ln -sf ~/agent-notes/bin/agent-notes /usr/local/bin/agent-notes
agent-notes install    # interactive wizard guides you through setup
agent-notes doctor

Update anytime with cd ~/agent-notes && git pull && agent-notes update.

What's Included

Component Description
Skills 31 on-demand knowledge modules (Rails, Docker, Git, Kamal)
Agents 14 specialized AI subagents with hierarchical model strategy
Rules Global instructions, code quality, and safety guardrails
Config Global instructions for Claude Code, OpenCode, and GitHub Copilot

CLI Reference

agent-notes <command> [options]
Command Description
install Interactive installer wizard (CLI, scope, skills selection)
install --local [--copy] Direct install to current project (no wizard)
uninstall [--local] Remove installed components
update Pull latest, rebuild, reinstall
doctor [--local] [--fix] Check installation health
info Show status and component counts
list [agents|skills|rules|all] List installed components
validate Lint source configuration files
memory [list|size|show|reset|export|import] [name] Manage agent memory

Supported platforms

Platform Install target Config format
Claude Code ~/.claude/ YAML frontmatter + Markdown prompts
OpenCode ~/.config/opencode/ YAML frontmatter + Markdown prompts
GitHub Copilot ~/.github/ copilot-instructions.md

Examples

# Interactive install (recommended)
agent-notes install

# Example wizard session:
#
#   Which CLI do you use?
#     1) [*] Claude Code
#     2) [*] OpenCode
#   Enter numbers to toggle (comma-separated), or press enter for defaults.
#   Choice:                          ← press enter to keep both
#
#   Where to install?
#     1) * Global (~/.claude, ~/.config/opencode)
#     2)   Local (current project)
#   Choice [1]:                      ← press enter for global
#
#   How to install?
#     1) * Symlink (auto-updates when source changes)
#     2)   Copy (standalone, allows local customization)
#   Choice [1]:                      ← press enter for symlink
#
#   Which skills to include?
#     1) [*] Rails — models, controllers, views, routes, testing (22 skills)
#     2) [*] Docker — Dockerfile, Compose patterns (4 skills)
#     3) [*] Kamal — deployment with Kamal (1 skill)
#     4) [*] Git — commit workflow, conventional commits (1 skill)
#   Choice:                          ← press enter for all
#
#   Ready to install:
#     CLI:      Claude Code + OpenCode
#     Scope:    Global (~/.claude, ~/.config/opencode)
#     Mode:     Symlink
#     Skills:   Rails (22), Docker (4), Kamal (1), Git (1)
#     Agents:   13 (Claude Code) + 13 (OpenCode)
#     Config:   CLAUDE.md, AGENTS.md
#     Rules:    2
#   Proceed? [Y/n]: Y

# Direct install (scripted, no wizard)
agent-notes install --local
agent-notes install --local --copy

# Check health and fix issues
agent-notes doctor --fix

# Manage agent memory
agent-notes memory list
agent-notes memory show coder
agent-notes memory reset reviewer

Agent Team

Specialized subagents with hierarchical model strategy: Opus 4.6 decides, Sonnet 4 executes, Haiku 4.5 explores.

Agent roster

Agent Model Role
lead Opus 4.6 Plans, delegates, reviews. The only Opus agent.
coder Sonnet 4 Implements features, fixes bugs, edits files.
reviewer Sonnet 4 Code quality review. Read-only.
security-auditor Sonnet 4 Security vulnerability analysis. Read-only.
test-writer Sonnet 4 Writes tests for any framework.
test-runner Sonnet 4 Diagnoses and fixes failing tests.
system-auditor Sonnet 4 Codebase health: duplication, N+1, coupling. Read-only.
database-specialist Sonnet 4 Schema design, indexes, query performance, migrations. Read-only.
performance-profiler Sonnet 4 Response times, memory, caching, bundle size. Read-only.
api-reviewer Haiku 4.5 API design, versioning, error handling, backward compatibility. Read-only.
tech-writer Haiku 4.5 Documentation: READMEs, API docs, changelogs.
devops Sonnet 4 Docker, CI/CD, deployment configs.
explorer Haiku 4.5 Fast file discovery and pattern search. Read-only.

4-phase lead workflow

1. ANALYZE — Lead reviews requirements, explores codebase
2. EXECUTE — Delegates to specialized agents (parallel execution)
3. REVIEW — Quality check by reviewer agents
4. VERIFY — Final validation and integration

Team diagram

You (human)
  |
  +-- Simple task ------> Main session (direct work)
  |
  +-- Complex task -----> Lead (Opus 4.6)
                           +-- Explorer (Haiku 4.5)     quick lookups
                           +-- Coder (Sonnet 4)         implementation
                           +-- Reviewer (Sonnet 4)      code review
                            +-- Test Writer (Sonnet 4)   tests
                            +-- Test Runner (Sonnet 4)   fix tests
                           +-- Security (Sonnet 4)      security audit
                           +-- Auditor (Sonnet 4)       codebase health
                           +-- DB Specialist (Sonnet 4) schema & queries
                           +-- Perf Profiler (Sonnet 4) performance
                           +-- API Reviewer (Haiku 4.5) API design
                           +-- Tech Writer (Haiku 4.5)  documentation
                           +-- DevOps (Sonnet 4)        infrastructure

Architecture

Single source of truth: agent_notes/data/buildagent_notes/dist/install

  1. Source — YAML metadata + Markdown prompts
  2. Build — Generate platform-specific configs
  3. Dist — Built artifacts ready for installation
  4. Install — Deploy via symlinks or copy

Project Structure

agent-notes/
├── bin/agent-notes          # CLI wrapper (entry point)
├── agent_notes/             # Python implementation
│   ├── __init__.py, cli.py  # Core modules
│   ├── VERSION              # Package version
│   ├── data/                # Single source of truth
│   │   ├── agents/
│   │   │   ├── agents.yaml  # Agent metadata
│   │   │   └── *.md         # Agent prompt files
│   │   ├── skills/          # Skill directories
│   │   ├── rules/           # Code quality rules
│   │   └── global.md        # Global instructions
│   └── dist/                # Built artifacts
│       ├── claude/, opencode/
│       ├── rules/
│       └── skills/
├── scripts/                 # Build/utility scripts
└── tests/                   # Test suite

Install Methods

Git clone + wrapper (recommended)

git clone https://github.com/rubakas/agent-notes.git ~/agent-notes
ln -sf ~/agent-notes/bin/agent-notes /usr/local/bin/agent-notes
agent-notes install

After cloning, agent-notes install launches the interactive wizard to configure your setup.

Pip install (future)

pip install agent-notes
agent-notes install

Project-Level Overrides

Use --local --copy for project-specific customizations. The wizard handles local installs too, while --local --copy is for scripted/CI use:

agent-notes install --local --copy

Then edit the copied files in .claude/ or .opencode/ directories.

Precedence: Project-level configs replace global versions entirely.

Skills

On-demand knowledge modules loaded mid-conversation.

Available skills

Rails: rails-models, rails-controllers, rails-routes, rails-concerns, rails-views, rails-views-advanced, rails-view-components, rails-view-components-advanced, rails-helpers, rails-javascript, rails-jobs, rails-mailers, rails-broadcasting, rails-migrations, rails-active-storage, rails-validations, rails-testing-controllers, rails-testing-models, rails-testing-system, rails-style, rails-controllers-advanced, rails-models-advanced, rails-initializers, rails-lib

Docker: docker-dockerfile, docker-dockerfile-languages, docker-compose, docker-compose-advanced

Kamal: rails-kamal

Git: git — git workflow, commit chunking, conventional commit messages

Usage

Claude Code / OpenCode:

Use the rails-models skill to help with this association
Load the docker-compose skill for multi-service setup

Development

Prerequisites

  • Python 3.9+
  • PyYAML (pip install pyyaml)

Running tests

python3 -m pytest tests/

Building

python3 -m agent_notes build

Validating

python3 -m agent_notes validate

Project layout

  • agent_notes/data/ — single source of truth (edit here)
  • agent_notes/dist/ — generated output (do not edit)
  • agent_notes/ — CLI implementation
  • tests/ — test suite
  • scripts/ — dev-only tools (release, etc.)

Contributing

When adding new content:

  1. Edit source files — all changes go in agent_notes/data/ directory
  2. Run buildagent-notes build to generate platform configs
  3. Validateagent-notes validate before committing
  4. Keep it generic — remove app-specific references
  5. Show examples — include code samples with explanations
  6. Stay modular — each skill should be independently usable
  7. Stay concise — agent prompts under 60 lines

License

MIT

About

Modular configuration files and best practices for AI coding agents across different technologies and frameworks.

Resources

License

Stars

Watchers

Forks

Contributors