convez.ai
MCP ServerAutomaçãoAvançado
appautomaton

Agent Designer

appautomaton/agent-designer

Skills portáteis e colaboração entre agentes com workflows estruturados

Ver no GitHub
BomPluginScore · de 0 a 100

Espaço de trabalho portátil para skills reutilizáveis em agentes de IA (Claude Code, Codex, Gemini). Estrutura workflows issue-driven com colaboração entre agentes via bridge scripts e continuidade de sessão, permitindo que múltiplos agentes trabalhem metodicamente em projetos.

Instalação

$npx convez-ai appautomaton/agent-designer

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
133
Forks
17
Acompanhando
1
Saúde
25%
Último commit
há 2 meses
Versão
—
Issues abertas
1
Contribuidores
1

Casos de uso

  • Coordenar trabalho entre Claude Code, Codex e Gemini em um projeto
  • Implementar workflows issue-driven com rastreamento automático de status
  • Criar skills reutilizáveis que normalizam comportamento de agentes
  • Delegar tarefas entre agentes com contexto persistente de sessão

Veredito

Ferramenta robusta e bem documentada para orquestração de múltiplos agentes de IA com colaboração estruturada. Ideal para equipes que precisam de workflows determinísticos entre diferentes LLMs, mas exige familiaridade com CLI tools e Python para explorar todo o potencial.

README

A tradução para o português deste README ainda está na fila.

Language: English | 中文

Agent Designer

A portable skills workspace for AI coding agents — Codex, Claude Code, and Gemini.

Design structured, reusable skills that give your agents clear workflows, safe defaults, and multi-turn collaboration capabilities. Clone this repo as a starting point for any project where you want agents to work methodically rather than freestyle.

What's inside

Collaboration skills — let your agents delegate to each other via bridge scripts with session continuity:

SkillPurpose
collaborating-with-claudeDelegate to Claude Code CLI (review, diff, consultation)
collaborating-with-geminiDelegate to Gemini CLI (review, web search, image analysis)
collaborating-with-codexDelegate to Codex CLI (implementation, diagnosis, review)

Issue-driven development — structure work into plans and trackable Issue CSVs:

SkillPurpose
issue-driven-workflowPlan → Issue CSV → autonomous execution with status tracking

Getting started

[!TIP]

  1. (Optional) Use this repo as a template: clone it, remove git history (rm -rf .git), re-init (git init).
  2. Write your AGENTS.md — project role, constraints, stack, safety rules.
    • To add issue-driven workflow: ask your agent to "apply AGENTS.issues.template.md on top of AGENTS.md".
  3. Start working — ask your agent to create a plan, generate an Issue CSV, or collaborate with another agent.
    • Example: "Create a plan and Issue CSV for ."

Project structure

skills/                          ← skill source (the real content)
  collaborating-with-claude/     ← bridge script + SKILL.md + references
  collaborating-with-gemini/     ← bridge script + SKILL.md + references
  collaborating-with-codex/      ← bridge script + SKILL.md + references + prompt recipes
  issue-driven-workflow/              ← plan/CSV workflow + templates + scripts
.codex/skills/                   ← symlinks (Codex wiring)
AGENTS.md                        ← project-specific rules
AGENTS.issues.template.md        ← issue-driven workflow (apply on top of AGENTS.md)

How it works

Each skill follows progressive disclosure:

  • Level 1 (metadata) — name + description in YAML frontmatter. Always loaded. Used for discovery.
  • Level 2 (instructions) — the SKILL.md body. Loaded when the skill is invoked. Contains workflow, safety, quick start.
  • Level 3 (resources) — scripts/, references/, assets/. Loaded on demand. Contains deep docs, templates, helper scripts.

Notes

  • This repo is workflow-first — most changes are text, not code.
  • Bridge scripts wrap CLI tools and return structured JSON with session continuity.
  • Skills are portable across Codex and Claude Code with minimal adaptation.
  • If invoking bridge scripts directly, prefer python3.

Acknowledgements

Inspired by:

Ver o original no GitHub

Bom para

Relacionados