Execute qualquer Skill no Manus
com um clique

Execute qualquer Skill no Manus com um clique

Começar

team-implement

Estrelas0

Forks0

Atualizado13 de junho de 2026 às 15:52

Spin up an agent team for a task with structured workflow. Provide a ticket reference, task description, or both.

Instalação

Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.

Executar no Manus

Fonte

CJHwong

CJHwong/claude-skills

Abrir repositório GitHub Ver repositórios do creator

Download

Executar no Manus

Ocupações relacionadasSOC

Baseado na classificação ocupacional SOC

Desenvolvedores de softwareInformática e Matemática·SOC 15-1252

SKILL.md

readonly

name	team-implement
description	Spin up an agent team for a task with structured workflow. Provide a ticket reference, task description, or both.
user-invocable	true
argument-hint	<ticket reference and/or task description>

Team Implement

Create an agent team for a task. Provide a ticket reference, task description, or both.

Workflow (PRAR)

The team follows a phased workflow. Do NOT skip phases.

Phase 1: PERCEIVE (Lead only)

If a ticket reference is provided, fetch it for full context
Read AGENTS.md at repo root and in relevant directories
Identify the scope of work and split into parallel-safe pieces
If requirements are vague, ask the user for clarity before proceeding

Phase 2: SCAFFOLD (Lead only)

Use spawn_agent to create each teammate
Each teammate's prompt must include:
- File paths they own (no overlap between teammates)
- Architecture context and interfaces they need to implement or consume
- Existing files they MUST read before planning
- Instructions to present plan, then wait for approval before coding

Phase 3: REASON (Lead + Teammates)

Teammates investigate their area and submit plans
Lead reviews plans and approves or rejects
If a teammate raises a blocking question, lead investigates before answering
Lead presents the combined plan to the user and waits for approval
After user approval, teammates proceed to coding

Phase 4: ACT (Teammates)

Each teammate writes tests alongside implementation
Each teammate runs tests on their own work before marking tasks complete
If a teammate gets stuck after 2-3 attempts, message the lead
Lead monitors progress and redirects as needed. Lead does NOT implement tasks
Teammates coordinate when their work produces shared interfaces

Phase 5: REFINE (Lead)

Wait for ALL teammates to finish
Verify no file conflicts between teammates
Run the full test suite
Present a summary to the user of what was done

Phase 6: CLEANUP (Lead)

Close all teammates

Team structure

Default to 3 teammates unless the task clearly needs more or fewer
Each teammate gets a clear role name (e.g., "backend", "frontend", "tests")

Rules

Do NOT push, create PRs, or run git operations beyond what's needed for implementation
Do NOT modify files outside the teammate's assigned scope
Minimal changes only. No refactoring beyond what the task requires
Stick to the repo's existing stack and patterns

Mais deste repositório

mesmo repositório

team-research

CJHwong/claude-skills

Spin up a research agent team to explore a topic from multiple angles. No code changes.

2026-06-130

git-commit

CJHwong/claude-skills

Use this skill when the user wants to turn current uncommitted local changes into one or more commits: "commit my changes", "make a commit", "split this work into commits", "organize staged/unstaged changes", "write commit messages", or "use [why]/[how] commit format". Inspect git status and diffs, group files by logical change, propose a commit plan, and wait for approval before staging or committing. Do not use for rewriting existing commits, rebasing, pushing, creating PRs, or remote GitHub operations.

2026-06-130

rewrite-commits

CJHwong/claude-skills

Use this skill when the user wants to clean up commits that already exist on the current branch: "squash commits", "rewrite history", "reword commit messages", "fold fixup commits", "autosquash", "clean up WIP commits", "organize branch history", or "prepare this branch for PR". Compare against the base branch, propose a safe rewrite plan, and wait for approval before rebasing or resetting. Do not use for uncommitted-only changes, creating new work commits, PR comments, CI checks, or GitHub remote operations.

2026-06-130

review-technical-claims

CJHwong/claude-skills

Audit a document for unverified technical claims, verify attributions against real sources, fix errors, and provide honest critical review.

2026-06-130

ollama-vision

CJHwong/claude-skills

Use this skill whenever the user asks you to look at an image on their machine — extract text (OCR), describe a screenshot or photo, read a diagram or chart, or answer questions about visual content. Trigger even if the user doesn't explicitly say "vision" or "OCR" — phrases like "what's in this screenshot", "read this image", "transcribe this photo", "what does this diagram show", or "can you see what's in this picture" all qualify. Also trigger when the user references an image file (.png, .jpg, .webp, .gif) with information they need extracted.

2026-06-130

prune-projects

CJHwong/claude-skills

Find stale .claude/projects/ folders whose source directories no longer exist on the filesystem, then clean them up (merge renamed, delete dead).

2026-06-130

name	team-implement
description	Spin up an agent team for a task with structured workflow. Provide a ticket reference, task description, or both.
user-invocable	true
argument-hint	<ticket reference and/or task description>

Team Implement

Create an agent team for a task. Provide a ticket reference, task description, or both.

Workflow (PRAR)

The team follows a phased workflow. Do NOT skip phases.

Phase 1: PERCEIVE (Lead only)

If a ticket reference is provided, fetch it for full context
Read AGENTS.md at repo root and in relevant directories
Identify the scope of work and split into parallel-safe pieces
If requirements are vague, ask the user for clarity before proceeding

Phase 2: SCAFFOLD (Lead only)

Use spawn_agent to create each teammate
Each teammate's prompt must include:
- File paths they own (no overlap between teammates)
- Architecture context and interfaces they need to implement or consume
- Existing files they MUST read before planning
- Instructions to present plan, then wait for approval before coding

Phase 3: REASON (Lead + Teammates)

Teammates investigate their area and submit plans
Lead reviews plans and approves or rejects
If a teammate raises a blocking question, lead investigates before answering
Lead presents the combined plan to the user and waits for approval
After user approval, teammates proceed to coding

Phase 4: ACT (Teammates)

Each teammate writes tests alongside implementation
Each teammate runs tests on their own work before marking tasks complete
If a teammate gets stuck after 2-3 attempts, message the lead
Lead monitors progress and redirects as needed. Lead does NOT implement tasks
Teammates coordinate when their work produces shared interfaces

Phase 5: REFINE (Lead)

Wait for ALL teammates to finish
Verify no file conflicts between teammates
Run the full test suite
Present a summary to the user of what was done

Phase 6: CLEANUP (Lead)

Close all teammates

Team structure

Default to 3 teammates unless the task clearly needs more or fewer
Each teammate gets a clear role name (e.g., "backend", "frontend", "tests")

Rules

Do NOT push, create PRs, or run git operations beyond what's needed for implementation
Do NOT modify files outside the teammate's assigned scope
Minimal changes only. No refactoring beyond what the task requires
Stick to the repo's existing stack and patterns