Перейти к содержимому
TakoAPI
Назад к навыкам

teamwork

от chenxinbest

Browser & Automation
816Число загрузок по данным clawskills.sh
2Звёзды
0Просмотры
0 лайковВойдите, чтобы поставить лайк навыку
—Ещё нет оценок
Ваша оценка

Кратко

Dynamically creates and manages AI agent teams for complex tasks.

Описание

Что делает этот навык

Manages a team of AI models to execute complex engineering tasks by assigning specialized roles, coordinating work through a central herald model, and tracking performance scores and API costs across providers. Supports OpenAI, Anthropic, and other providers with three pricing models: pay-per-use, subscription, and tiered usage. Roles include architect, developer, tester, reviewer, analyst, and project manager.

Rather than manually switching between models and tracking their outputs, the skill automates role assignment, cost accounting, and inter-model coordination, with peer-evaluated performance scores that improve future role assignments over time.

Когда его использовать

  • Building a full-stack app with separate models handling architecture and implementation
  • Running parallel code review and testing across multiple AI providers
  • Assigning cost-optimized models to routine subtasks while reserving capable models for complex reasoning
  • Executing a multi-phase data pipeline design with structured handoffs between roles
  • Managing monthly AI API budgets across providers during coordinated team tasks

Пример сценария

ВВОД

User asks: Build a REST API for user authentication

АГЕНТ
  1. 1

    Decomposes the request into phases (analysis, architecture, implementation, testing, review) and presents the breakdown for user confirmation

  2. 2

    Queries configured providers, filters available models by status and capabilities, and convenes a team meeting to define required roles

  3. 3

    Assigns roles to models using a combined score of capability, cost efficiency, and current availability, then selects the fastest model as herald

  4. 4

    Models execute assigned tasks in parallel while the herald relays progress updates, detects blockers, and reports status to the host model

  5. 5

    Convenes a review meeting, aggregates outputs, updates peer performance scores, and produces a task report with cost breakdown

ВЫВОД

Complete REST API implementation with task execution report, per-model cost breakdown, and updated capability scores stored for future role assignments

Требования

  • API key for each AI provider you configure (e.g., OPENAI_API_KEY, ANTHROPIC_API_KEY)

Установка

clawhub install chenxinbest/teamwork

Выполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.

Ссылки