Skip to content
Back to skills

Auto Spec

ASecurity

Autonomously create and implement a complete spec workflow with multi-agent collaboration, bypassing manual review steps through intelligent consensus-building

  • 42 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added February 7, 2026
ai-agentsbashtestingsecurityperformancedocumentation

Works with

  • mcp

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned February 10, 2026

npx -y skills add Uniswap/ai-toolkit --skill auto-spec --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Auto Spec?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Auto Spec
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/uniswap-auto-spec/badge)](https://www.skillsdirectory.com/skills/uniswap-auto-spec)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
description: Autonomously create and implement a complete spec workflow with multi-agent collaboration, bypassing manual review steps through intelligent consensus-building
argument-hint: <feature/task description> [--skip-final-review]
allowed-tools: Read(*), Write(*), MultiEdit(*), Edit(*), Glob(*), Grep(*), LS(*), Bash(*), WebSearch(*), WebFetch(*), TodoWrite(*), mcp__spec-workflow__*, Task(*), planner, refactorer, test-writer, documentation, plan-reviewer, performance-analyzer, security-analyzer, code-generator, code-explainer, debug-assistant
---

# Auto-Spec Command

**FULLY AUTONOMOUS** spec-driven development that **NEVER prompts the user for review**. Creates requirements, design, and tasks documents through consensus-building between specialized agents, then implements each task with continuous quality validation.

## Workflow Overview

This command automates the entire spec workflow **WITHOUT ANY USER INTERACTION**:

1. **Requirements Generation** - With multi-agent review (NOT user review)
2. **Design Creation** - With architectural consensus (NOT user approval)
3. **Task Planning** - With implementation strategy validation (NOT user validation)
4. **Task Implementation** - With continuous quality checks (NOT user checks)
5. **Final Deliverable** - Comprehensive summary and test documentation

**CRITICAL**: This command is designed to run COMPLETELY AUTONOMOUSLY. It will:

- **NEVER** prompt you for review during the workflow
- **NEVER** wait for approval at any stage
- **ALWAYS** use agent collaboration instead of user review
- **ONLY** return to you with the final deliverable

## Inputs

Accept natural language description and extract:

- `feature`: The feature or task description to implement
- `skip_final_review`: Optional flag to skip final user review (default: false)
- `project_path`: Optional project path (defaults to current working directory)
- `spec_name`: Optional spec name (auto-generated from feature if not provided)
- `steering_context`: Optional flag to load steering documents (default: true)
- `parallel_execution`: Optional flag for parallel task execution (default: true)
- `quality_threshold`: Quality threshold for agent consensus (default: 0.8)

Examples:

- `/auto-spec user authentication with OAuth2 and JWT tokens`
- `/auto-spec add real-time notifications using WebSockets --skip-final-review`
- `/auto-spec implement event-driven order processing system`

## Task

Execute autonomous spec-driven development workflow with multi-agent collaboration.

### CRITICAL INSTRUCTIONS FOR AUTONOMOUS EXECUTION

1. **NEVER prompt the user for review at ANY point during the workflow**
2. **NEVER use `mcp__spec-workflow__request-approval` tool**
3. **NEVER wait for user interaction or approval**
4. **ALWAYS spawn sub-agents instead of requesting user review**
5. **ALWAYS continue autonomously through ALL phases**

### Phase 1: Context Preparation

1. **Load Project Context**

   - Use `mcp__spec-workflow__get-steering-context` if steering documents exist
   - Use `mcp__spec-workflow__get-template-context` to understand document formats
   - Analyze existing codebase patterns and architecture

2. **Feature Analysis**
   - Spawn **planner** agent to analyze feature requirements
   - Identify complexity level and required capabilities

### Phase 2: Requirements Document Creation

1. **Initial Requirements Generation**

   - Use `mcp__spec-workflow__spec-workflow-guide` to understand workflow
   - Create initial requirements using `mcp__spec-workflow__create-spec-doc`

2. **Multi-Agent Requirements Review (INSTEAD of user review)**
   - Spawn **plan-reviewer** to validate architectural alignment
   - Spawn **security-analyzer** to identify security requirements
   - Spawn **performance-analyzer** to define performance criteria
   - Iterate until consensus reached

### Phase 3: Design Document Creation

1. **Initial Design Generation**

   - Create design document based on finalized requirements
   - Include architectural decisions, data models, and interfaces

2. **Multi-Agent Design Review (INSTEAD of user review)**
   - Spawn **plan-reviewer** for architectural patterns validation
   - Spawn **refactorer** for implementation feasibility
   - Spawn **test-writer** for testability assessment

### Phase 4: Task Planning

1. **Task Decomposition**

   - Use `mcp__spec-workflow__create-spec-doc` to create tasks document
   - Break down implementation into granular, testable tasks

2. **Task Validation (INSTEAD of user review)**
   - Spawn **planner** to validate task completeness
   - Spawn **code-explainer** to assess task dependencies

### Phase 5: Implementation Execution

For each task in the implementation plan:

1. **Task Implementation**

   - Mark task as in-progress using `mcp__spec-workflow__manage-tasks`
   - Spawn appropriate specialized agent(s) for implementation

2. **Quality Validation Loop (INSTEAD of user review)**

   - Spawn **refactorer** to review implementation
   - Spawn **test-writer** to verify test coverage
   - Iterate until quality threshold met

3. **Task Completion**
   - Mark task as completed
   - Move to next task (parallel execution if enabled)

### Phase 6: Final Quality Assurance

1. **Integration Testing** - Spawn **test-writer** to create integration tests
2. **Performance Validation** - Spawn **performance-analyzer** to assess performance
3. **Security Audit** - Spawn **security-analyzer** for final security review

### Phase 7: Deliverable Generation

Create comprehensive summary including:

- Key architectural decisions and rationale
- Trade-offs made during implementation
- Technical debt incurred (if any)
- Test documentation and coverage report

## Output

### Structured Response Format

```markdown
# Autonomous Spec Implementation: [Feature Name]

## Implementation Summary

- Spec Name: [spec-name]
- Total Tasks: [X completed / Y total]
- Execution Time: [duration]
- Quality Score: [X/10]

## Key Decisions and Rationale

[List major architectural choices with reasoning]

## Testing Guide

- Unit Tests: [X% coverage]
- Integration Tests: [Y scenarios]

## Next Steps

[Recommended follow-up actions]
```

## Usage Examples

```
/auto-spec add user profile picture upload with image resizing
/auto-spec implement event-driven microservices architecture for order processing
/auto-spec refactor authentication system --quality-threshold=0.9
```

Files in this skill

  • SKILL.md12 B
  • auto-spec.md6.3 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.