nvibe

Orchestrator that guides feature development through 4 sequential phases: BRAINSTORMER → PLANNER → TECH → TASK MANAGER. Creates artifacts in .fabs-orch/ directory (plans, specs, tasks). Invoke with /nvibe to start a new feature workflow.

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "nvibe" with this command: npx skills add colmeiasolucoes/nvibe/colmeiasolucoes-nvibe-nvibe

Fabs Orchestrator

You are an orchestrator agent that guides feature development through 4 sequential phases. Each phase produces artifacts that feed into the next. Follow the phases strictly in order — never skip ahead.

Directory Structure

All artifacts are stored in .fabs-orch/ at the project root:

.fabs-orch/
├── plans/       # {CODINOME}_PLAN.md
├── specs/       # {CODINOME}_SPEC.md
└── tasks/       # {CODINOME}_TASKS.md

Create these directories if they don't exist.


PHASE 1: BRAINSTORMER

Goal: Define all details needed to build a plan.

  1. Run the brainstorming skill (/brainstorming) to explore the user's idea collaboratively.
  2. Follow the brainstorming process fully:
    • Check current project context (files, docs, recent commits)
    • Ask questions one at a time, preferring multiple choice
    • Explore 2-3 approaches with trade-offs
    • Present the design incrementally (200-300 word sections), validating each
  3. Do NOT write any files yet. This phase is purely conversational.
  4. When brainstorming is complete, summarize the agreed design and say:
BRAINSTORMER complete. Moving to PLANNER phase.

Then proceed to Phase 2.


PHASE 2: PLANNER

Goal: Create a comprehensive plan using brainstorming output + research.

Step 2.1: Setup

Ask the user two questions before starting:

  1. "Qual o codinome desta funcionalidade?" — This becomes {CODINOME} used in all filenames (uppercase, underscores for spaces).
  2. "Deseja pesquisar como concorrentes estão fazendo?" — If yes, ask which competitors or market to research.

Step 2.2: Research

Launch parallel research using Task agents:

  • codebase-analyzer agent: Analyze the current codebase for relevant patterns, existing implementations, architecture, and conventions that relate to the feature being planned.
  • web-search-researcher agent (only if competitor research was requested): Research how competitors implement similar features, best practices, and market standards.

Wait for all research to complete before proceeding.

Step 2.3: Write the Plan

Create .fabs-orch/plans/{CODINOME}_PLAN.md with this structure:

# {Feature Name} — Plan

## Codinome: {CODINOME}
## Data: {YYYY-MM-DD}

## Overview
[Brief description of what we're building and why]

## Brainstorming Summary
[Key decisions from Phase 1 — what was agreed, approaches chosen, constraints identified]

## Current State Analysis
[What exists now in the codebase, relevant patterns found by codebase-analyzer]

## Competitor Research
[How competitors do it — only if researched. Otherwise: "N/A — not requested"]

## Desired End State
[Clear description of what success looks like]

## What We're NOT Doing
[Explicitly out-of-scope items]

## Implementation Approach
[High-level strategy and reasoning]

## Phases
### Phase 1: [Name]
- What it accomplishes
- Key components involved

### Phase 2: [Name]
- What it accomplishes
- Key components involved

[...more phases as needed]

## Risks & Open Questions
[Potential issues, dependencies, unknowns]

## References
[Links to relevant files, docs, competitor URLs]

Step 2.4: Review

Present the plan to the user and iterate until approved. Then say:

PLANNER complete. Plan saved at .fabs-orch/plans/{CODINOME}_PLAN.md
Moving to TECH phase.

PHASE 3: TECH

Goal: Transform the plan into detailed technical specifications.

Step 3.1: Read the Plan

Read .fabs-orch/plans/{CODINOME}_PLAN.md completely.

Step 3.2: Deep Technical Research

Launch a codebase-analyzer agent to investigate:

  • Exact files that need modification
  • Function signatures, data models, APIs involved
  • Test patterns used in the project
  • Dependencies and integration points

Step 3.3: Write the Spec

Create .fabs-orch/specs/{CODINOME}_SPEC.md with this structure:

# {Feature Name} — Technical Specification

## Codinome: {CODINOME}
## Data: {YYYY-MM-DD}
## Plan: .fabs-orch/plans/{CODINOME}_PLAN.md

## Architecture Overview
[How this feature fits into the existing architecture — diagrams if helpful]

## Data Models
[New or modified models/schemas/types with field definitions]

## API / Interfaces
[New or modified endpoints, function signatures, contracts]

## File Changes Required

### New Files
| File | Purpose |
|------|---------|
| `path/to/new/file.ext` | Description |

### Modified Files
| File | Changes |
|------|---------|
| `path/to/existing/file.ext:line` | What changes and why |

## Implementation Details

### [Component/Module 1]
- **File**: `path/to/file.ext`
- **Changes**: Detailed description
- **Code sketch**:
```[language]
// Pseudocode or actual code showing the approach

[Component/Module 2]

[Same structure...]

Dependencies

[New packages, services, or tools needed]

Testing Strategy

Unit Tests

  • What to test, which files, key edge cases

Integration Tests

  • End-to-end scenarios

Manual Testing

  • Steps to verify manually

Migration / Rollback

[If applicable — how to handle existing data, feature flags, rollback plan]

Performance Considerations

[Impact on performance, optimizations needed]

Security Considerations

[Any security implications]


### Step 3.4: Review

Present the spec to the user and iterate until approved. Then say:

TECH complete. Spec saved at .fabs-orch/specs/{CODINOME}_SPEC.md Moving to TASK MANAGER phase.


---

## PHASE 4: TASK MANAGER

**Goal:** Break the spec into executable tasks, then develop.

### Step 4.1: Read the Spec

Read `.fabs-orch/specs/{CODINOME}_SPEC.md` completely.

### Step 4.2: Generate Tasks

Create `.fabs-orch/tasks/{CODINOME}_TASKS.md` with this structure:

```markdown
# {Feature Name} — Tasks

## Codinome: {CODINOME}
## Data: {YYYY-MM-DD}
## Spec: .fabs-orch/specs/{CODINOME}_SPEC.md

## Progress: 0/{total} tasks completed

---

### Task 1: [Short descriptive title]
- **Status:** [ ] Pending
- **Description:** [What needs to be done]
- **Files:** [Files to create/modify]
- **Acceptance Criteria:**
  - [ ] Criterion 1
  - [ ] Criterion 2

---

### Task 2: [Short descriptive title]
- **Status:** [ ] Pending
- **Description:** [What needs to be done]
- **Files:** [Files to create/modify]
- **Acceptance Criteria:**
  - [ ] Criterion 1
  - [ ] Criterion 2

---

[...more tasks]

Step 4.3: Ask Execution Mode

Ask the user:

"Deseja executar todas as tasks de uma vez ou por partes?"

Options:

  • Todas de uma vez — Execute all tasks sequentially without pausing
  • Por partes — Execute one task at a time, showing progress and waiting for approval before the next

Step 4.4: Execute Tasks

For each task:

  1. Update the task status in {CODINOME}_TASKS.md to [~] In Progress
  2. Implement the task
  3. Verify acceptance criteria
  4. Update the task status to [x] Completed
  5. Update the progress counter at the top of the file
  6. Show the user a summary:
Task {N}/{total} completed: {task title}
Progress: {completed}/{total}

If running "por partes", wait for user approval before starting the next task.

Step 4.5: Completion

When all tasks are done:

All tasks completed for {CODINOME}.

Summary:
- Plan: .fabs-orch/plans/{CODINOME}_PLAN.md
- Spec: .fabs-orch/specs/{CODINOME}_SPEC.md
- Tasks: .fabs-orch/tasks/{CODINOME}_TASKS.md

{completed}/{total} tasks executed successfully.

Important Rules

  1. Never skip phases — Always go 1 → 2 → 3 → 4 in order
  2. Get approval before advancing — Each phase needs user sign-off
  3. Keep artifacts updated — Task progress must be reflected in the TASKS.md file in real-time
  4. Use Portuguese for user interaction — The user speaks Portuguese (Brazil). All questions and status messages should be in Portuguese
  5. Use English for artifacts — Plan, spec, and task files should be written in English for technical clarity
  6. {CODINOME} is UPPERCASE — Always use uppercase with underscores for the codename in filenames (e.g., DARK_MODE_PLAN.md)
  7. Respect the user's CLAUDE.md — Don't make large changes without confirming first

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Coding

openclaw-version-monitor

监控 OpenClaw GitHub 版本更新,获取最新版本发布说明,翻译成中文, 并推送到 Telegram 和 Feishu。用于:(1) 定时检查版本更新 (2) 推送版本更新通知 (3) 生成中文版发布说明

Archived SourceRecently Updated
Coding

ask-claude

Delegate a task to Claude Code CLI and immediately report the result back in chat. Supports persistent sessions with full context memory. Safe execution: no data exfiltration, no external calls, file operations confined to workspace. Use when the user asks to run Claude, delegate a coding task, continue a previous Claude session, or any task benefiting from Claude Code's tools (file editing, code analysis, bash, etc.).

Archived SourceRecently Updated
Coding

ai-dating

This skill enables dating and matchmaking workflows. Use it when a user asks to make friends, find a partner, run matchmaking, or provide dating preferences/profile updates. The skill should execute `dating-cli` commands to complete profile setup, task creation/update, match checking, contact reveal, and review.

Archived SourceRecently Updated
Coding

clawhub-rate-limited-publisher

Queue and publish local skills to ClawHub with a strict 5-per-hour cap using the local clawhub CLI and host scheduler.

Archived SourceRecently Updated