DEVELOPER TOOLS

Ship with AI Agents

Production-tested tools for multi-agent orchestration, image editing, and persistent memory. Born from running 10 agents across 3,100+ tasks.

open source Free

Agent Architect Kit

The complete starter kit for building multi-agent systems with Claude Code. Battle-tested configs, agent definitions, and process docs extracted from a production system.

$ tree agent-architect-kit/
├── CLAUDE.md.template
├── agents/
│ ├── coder.md
│ ├── qa.md
│ ├── designer.md
│ ├── operations.md
│ ├── marketing.md
│ └── product.md
├── memory/
│ └── directive.md
└── processes/
├── work-queue.md
├── orchestration.md
└── 10 more...
Battle-tested CLAUDE.md template with annotated rules
6 agent role definitions (coder, QA, designer, ops, marketing, product)
Memory protocol + agent directive system
10 process docs + orchestration scripts
Incident learnings database + customization guide
Get the Kit — Free
open source Free

Agent Orchestra CLI

Open-source CLI that orchestrates multiple Claude Code agents from a task queue. Define roles, queue work, let agents claim and complete tasks autonomously.

$ orchestra init
Created agents.yml with 3 starter roles
$ orchestra add coder "Add user auth"
Task #1 added (role: coder)
$ orchestra run
â–¸ Spawning coder agent for task #1...
â–¸ Agent claimed task, working...
✓ Task #1 complete (2m 34s)
✓ Zero framework dependency — pure Ruby CLI
✓ YAML-based task queue — no database required
✓ Configurable agent roles with tool restrictions
✓ Built-in health monitoring + stale task recovery
✓ Works with any Claude Code project
View on GitHub
open source Free

AgentBrush

Image editing toolkit built for AI agents. Remove backgrounds, composite layers, add text overlays, and validate images against built-in presets.

$ pip install agentbrush
Successfully installed agentbrush
>>> from agentbrush import canvas
>>> img = canvas.remove_bg("photo.png")
✓ Background removed (flood-fill)
>>> canvas.validate(img, preset="social")
✓ Passed: dimensions, format, DPI
Edge-based flood-fill background removal
Layer compositing with alpha blending
Text overlays with Pillow integration
Spec validation with built-in presets (social, icons, thumbnails)
Designed for headless / CI pipeline usage
View on GitHub
open source Free

Agent Cerebro

Persistent memory for AI agents. Two-tier storage with short-term markdown files and long-term SQLite + embeddings. Agents remember across sessions.

$ pip install agent-cerebro
Successfully installed agent-cerebro
>>> from agentrecall import memory
>>> memory.store("coder", "gotcha",
"SQLite has no ILIKE")
✓ Stored (dedup: 0.31 similarity)
>>> memory.search("case insensitive")
"SQLite has no ILIKE" (0.89)
Two-tier: short-term markdown + long-term SQLite
Semantic search via OpenAI embeddings
Auto-dedup at cosine similarity > 0.92
Role-scoped memory (coder, qa, designer, etc.)
CLI tools: memory-store, memory-search, memory-check
View on GitHub
PREMIUM

Production-Ready Kits

Everything you need to run AI agents in production. Instant download after purchase.

instant download $99.00

Rails AI CEO Template

The complete Rails 8 framework for running autonomous AI operations. Built from a production system that has processed 5,000+ tasks with 10 specialized agents. What's included: - Multi-agent orche...

Rails AI CEO Template
View Details
instant download $29.99

Agent Starter Kit Pro

The complete production-ready kit for building multi-agent systems with Claude Code. Includes 8 agent role definitions, memory protocol, orchestration scripts, process docs, and a battle-tested CLA...

Agent Starter Kit Pro
View Details

Built in Production

These tools were extracted from a live production system. Not a tutorial — real configs from real agent sessions running real tasks.

0
Open Source Tools
0
Agent Roles
0
Tasks Processed
0
Lines in CLAUDE.md

Battle-Tested

Every rule in CLAUDE.md exists because something broke without it. Incident dates and commit hashes included.

Production Hours

Agents run autonomously on queued tasks — coder, QA, designer, product, marketing, operations, and more.

Failure-Hardened

3-retry task failure handling, stale claim recovery, rate limit detection, OOM prevention — all from real incidents.

Ready to Orchestrate?

4 free tools for agent orchestration, image editing, and persistent memory. Grab the Kit to get started.