ES
SpecLeap
Open Source Framework

SpecLeap
Framework

Transform VSCode, Cursor, JetBrains, Vim, Zed, or any IDE with an AI assistant into a spec-first development machine. Immutable CONTRACT, specialized agents, and 34 professional Agent Skills.

View on GitHub → View Features

✅ 100% Free · ✅ Open Source (MIT) · ✅ 34 Agent Skills (20 TIER 1 + 14 TIER 2) · ✅ Spanish + English

0
Agent Skills
0
Max Questions
0
Validation Layers
0
Commands

❌ Without SpecLeap

  • • AI improvises code without context
  • • Incomplete specs → wrong implementation
  • • No memory → repeats mistakes
  • • Uncontrolled scope changes
  • • Generic and vulnerable code

✅ With SpecLeap

  • • CONTRACT defines immutable scope
  • • Context/ stores architectural decisions
  • • Specialized agents follow standards
  • • Conversational commands guide the flow
  • • 34 Agent Skills (20 TIER 1 + 14 TIER 2) apply best practices
What's included

Key
Features

01 · Killer Feature
34
TIER 1 + TIER 2

Agent Skills

Security (OWASP, STRIDE), professional design (Vercel-style), DRY consistency, TDD, Laravel/React best practices.

02

3-Mode Survey

MVP (58 questions), Serious Project (93 questions with governance, SLAs, compliance) or Creative Mode (describe in a paragraph and the AI extracts what you said). Spanish/English, smart validations, partial saves.

03

Immutable CONTRACT

Defines scope and never changes. Change proposals via refinar command or openspec CLI.

04

Conversational Commands

refinar planificar implementar documentar

Natural conversational workflow.

05

Specialized Agents

Backend, Frontend, Product. Each with domain-specific rules.

06

Asana Integration

Automatically generates backlog from CONTRACT.md. Epics, user stories and tasks ready for sprint.

07

Monorepo Multi-Project

Each independent project in proyectos/ with its CONTRACT, context/ and code.

08

Global Standards

PHP/Laravel, JavaScript/React, CSS/Tailwind, documentation. Consistency across all projects.

09

OpenSpec CLI

Formal workflow with change proposals, delta specs, verification and archiving. Perfect for enterprise teams.

10

Total Discoverability

Type ayuda in chat and see ALL available commands.

11 Featured

Automatic 3-Layer Validation

Pre-commit git hooks (<5 sec), full validation in implementar (1-5 min), CodeRabbit on PR (5-10 min). Quality guaranteed.

12 ES · EN

Multilingual

Entire framework translated: installer, survey, commands, documentation. Select your preferred language in the first step.

Adaptive survey

Choose how to
build your CONTRACT

Three modes depending on the type and maturity of your project. Start simple and grow later: you can switch modes at any time.

Mode 1 · Fast

MVP

Prototype / Side-project / Small team

  • → 58 structured questions
  • → 15-20 minutes
  • → Stack, core features, design, basic deploy, constraints
  • → Auto-save every 10 questions

Ideal when you already know the scope and want a CONTRACT in under 30 min.

Recommended
Mode 2 · Enterprise

Serious Project

Team + Departments + Compliance

  • → 93 questions (the 58 of MVP + 35 extras)
  • → 25-35 minutes
  • → Governance, departments involved, enterprise integrations (SSO, ERP, CRM, BI)
  • → SLAs (uptime, RTO, RPO, DR, on-call)
  • → Compliance (GDPR, HIPAA, PCI, SOC2, ISO27001)
  • → Roadmap by phases, deploy strategy (canary, blue-green), observability, multi-tenant/region/i18n

Ideal for enterprise products with multiple teams, regulatory requirements or committed SLAs.

Mode 3 · Conversational

Creative

You Tell · The AI Extracts

  • → Describe the project in one or several paragraphs
  • → 5-10 minutes
  • → The AI extracts stack, audience, features and constraints from your text
  • → Shows you a summary to validate, then asks only the critical gaps
  • → No hallucinations: if you don't say it, it asks you

Ideal when you already know what you want and prefer a conversation over a form.

🔄

You can switch between modes

Once the survey is done, SpecLeap offers to upgrade to Serious (add governance, compliance, SLAs without repeating what's already answered) or add Additional Context in Creative mode (a free-form paragraph appended to the CONTRACT). It's iterative: repeat as many times as needed until the CONTRACT reflects your project exactly.

Quality guaranteed

Validation in
3 Layers

SpecLeap automatically validates your code in 3 progressive levels before reaching production.

1
< 5 sec
Automatic

Pre-Commit Git Hook

Runs before every commit. Fast validation in less than 5 seconds.

Validates

  • → Syntax (PHP, JS/TS)
  • → Linters (ESLint, PHPStan)
  • → Formatters (Prettier, PHP-CS-Fixer)
  • → CONTRACT.md not modified
  • → No console.log / var_dump

Installation

bash scripts/install-git-hooks.sh

❌ If fails: Rejects commit

2
1-5 min
Pre-push

Full Validation in implementar

Runs when finishing a feature before push. Exhaustive validation.

Validates

  • → Unit tests
  • → Integration tests
  • → Compliance vs CONTRACT.md
  • → Compliance vs specs
  • → Coverage ≥ 80%

Result

✅ If passes: Commit + Push + Creates PR

❌ If fails: No push, shows errors

3
5-10 min
On PR

CodeRabbit Review

Runs automatically when creating PR on GitHub. Deep review.

Validates

  • → Architecture and patterns
  • → Security (vulnerabilities)
  • → Business logic
  • → Cyclomatic complexity
  • → Documentation

Setup

  1. 1. .coderabbit.yaml already included
  2. 2. Install GitHub app
  3. 3. Open PR → Reviews automatically
Result

Fewer bugs. Faster reviews.
Continuous learning.

All 3 layers work together to ensure only quality code reaches production.

The difference

34 Agent
Skills

Agent Skills are specialized extensions that transform the quality of AI-generated code.

20
Tier 1
+
14
Tier 2
✕
Before

Without Skills

  • — Generic Bootstrap
  • — Basic vulnerabilities
  • — Frequent code duplication
  • — Inconsistent APIs
  • — Coverage <60%
✓
After

With TIER 1 Skills

  • → Professional Vercel/Linear style
  • → OWASP Top 10 + STRIDE threat modeling
  • → Automatic DRY, verify before creating
  • → REST/GraphQL best practices
  • → TDD methodology, coverage ≥90%
Tier 1 breakdown

20 skills across 4 categories

5
🔒

Security

  • → SAST config
  • → STRIDE threat
  • → Security extraction
  • → Backend/API security
  • → Frontend/Mobile security
3
⚙

Consistency

  • → Verification-before-completion
  • → Code review excellence
  • → Systematic debugging
6
◆

Design

  • → Web design (Vercel)
  • → Frontend (Anthropic)
  • → UI/UX pro max
  • → Tailwind system
  • → shadcn/ui
  • → Responsive design
6
⚡

Backend

  • → Laravel specialist
  • → React best practices
  • → TDD
  • → API design
  • → PostgreSQL design
  • → Error handling
End-to-end

Complete
Workflow

From idea to working code in clear steps.

1
20 min

Run Survey

# Say "Hi" to Claude Code and pick a mode:
#   MVP        → 58 questions (15-20 min)
#   Serious    → 93 questions (25-35 min, governance + SLAs + compliance)
#   Creative   → describe in a paragraph, AI extracts and only asks gaps
#
# Auto-save every 10 questions

Generates complete CONTRACT.md with all project information. You can switch modes after generating it.

2
Mandatory

Push to GitHub (MANDATORY) 

git init
git add .
git commit -m "init: CONTRACT.md + context/"
git push origin main

# CONTRACT.md is your source of truth
# Must be versioned from the start

⚠️ CONTRACT must be on GitHub before continuing.

3
Asana

Plan in Asana

# In chat with Claude/Cursor:
planificar

# Auto-generates backlog from CONTRACT.md
# Creates sections + user stories + tasks
4
Code

Implement According to Spec

# In chat:
implementar "Google login"

# Verifies if it's in CONTRACT
# Reads context/ for context
# Generates code following standards
5
Docs

Document

# In chat:
documentar "User API"

# Generates OpenAPI/Swagger docs
# Updates technical README
Your IDE, your choice

Works with
any IDE

SpecLeap is IDE-agnostic. It works with any editor that has an AI assistant: VSCode, Cursor, JetBrains, Vim, Zed, or whatever you use.

VSCode ✦ Cursor ✦ JetBrains ✦ Vim ✦ Zed ✦ Copilot ✦ Continue ✦ Codeium ✦ Gemini ✦ Tabnine ✦

These are just a few examples. If your IDE has AI, SpecLeap works.

V

VSCode + Claude Code

Install the Claude Code extension in VSCode. SpecLeap works automatically with conversational commands and Agent Skills.

  • → Reads CLAUDE.md
  • → Commands: refinar, planificar, etc.
  • → Automatic Agent Skills
C

Cursor / Continue

Native IDEs with integrated Claude. Out-of-the-box full support.

  • → Native conversational commands
  • → Memory-bank in context/
  • → Integrated Agent Skills
G

GitHub Copilot

Uses .github/copilot-instructions.md + openspec CLI.

  • → Global standards in rules/
  • → Workflow via CLI
  • → Works in VSCode, JetBrains, Vim
★

Any IDE with AI

Gemini Code Assist, Codeium, Tabnine, or your preferred tool.

  • → Reads .md instruction files
  • → Standards in rules/
  • → openspec CLI always available
What matters

It doesn't matter which IDE you use. SpecLeap works because it reads files from your project (CONTRACT.md, rules/, context/), not because it depends on a specific tool.

Start in 10 min

Quick
Installation

Ready in 10-15 minutes. Choose your method: npm (faster) or GitHub (development).

Recommended
Method A

via npm

Automatic installation with a single command.

npx specleap-framework@latest

✅ Download + interactive installer

✅ Always the latest stable version

Method B

via GitHub

For contributors or local development.

git clone https://github.com/ConceptualCreative/specleap-framework.git
cd specleap-framework
bash install.sh

✅ Full control of source code

✅ Ideal for contributors

After installing

3 steps to start

Both methods follow this flow.

1
10-15 min

Adaptive Survey (3 modes)

When you say Hi, the AI asks what kind of project you're building and picks the right survey: MVP, Serious Project or Creative Mode.

What it generates

  • 1.Select language (ES/EN) and an optional offer for markitdown to reduce tokens when processing PDF/Word/Excel
  • 2.Creates folder structure for your project
  • 3.Asks for the mode and applies the right survey
  • 4.Complete CONTRACT.md + agents + conversational commands
  • 5.Lets you switch between modes afterwards: start simple and grow, or vice versa
2
Mandatory

Configure Quality System

Asana + 34 Agent Skills + CodeRabbit + Git Hooks. 

# Asana + 34 Agent Skills (20 TIER 1 + 14 TIER 2) + CodeRabbit + Git Hooks
bash scripts/setup-mcp.sh

# Or configure manually:
nano .env

Mandatory components

Asana CodeRabbit 34 Skills Git Hooks GitHub

Without these components SpecLeap is NOT professional.

See SETUP.md for details.

3
Ready!

Open in Your IDE

Open the project in your favorite AI-powered editor. 

code .        # VSCode
cursor .      # Cursor

# In the chat:
ayuda

Done! Type ayuda (help) to see all available commands.

Donations

Support the
Project

SpecLeap is 100% free and open-source. If you find it useful, you can support its development with a donation.

Support on Ko-fi →
Your support helps
  • 01 Keep the framework updated
  • 02 Develop new features
  • 03 Improve documentation
  • 04 Create more Agent Skills
  • 05 Support the community
Frequently asked

FAQ

The questions that come up when evaluating SpecLeap. If yours isn't here, write to info@specleap.com.

01 · Vendor lock-in

Am I locked into Anthropic / OpenAI / Google?

No. SpecLeap is provider-agnostic by construction, not as a later add-on. The framework makes no calls to any model API: zero traffic to api.anthropic.com, api.openai.com or any other provider.

All the project's "knowledge" lives in .md and .sh files inside your repo. Each assistant reads its own instruction file: Claude reads CLAUDE.md, Cursor reads .cursorrules, GitHub Copilot reads .github/copilot-instructions.md, and so on across 9+ supported assistants.

The specification is the asset, not the model that processes it.

02 · SDD vs vibe coding

Why SDD if "vibe coding" with AI is faster?

Vibe coding is fast at the start and expensive at the end. After 50-100 iterations the AI loses context, contradicts existing patterns, reinvents utilities. You become the long-term memory system.

SDD captures the project's intent once, in an immutable CONTRACT. The AI can iterate 1000 times but always returns to the same north. Vibe coding scales with developer time; SDD scales with spec quality.

03 · Project size

Is it for small projects or only enterprise?

There's a minimum where SpecLeap pays off: roughly one week of real work.

  • → Weekend side-project: probably not — the initial survey is overhead you won't recover.
  • → 1-4 week project: yes, clearly — a single "no, this isn't what I meant" cycle already costs more than the 20-min survey.
  • → Multi-month with several contributors: fundamental.
  • → Enterprise / regulated: especially — versioned spec traceability is what audits ask for.
04 · Without Claude Code

Does it work without Claude Code? What do I lose?

Yes, it works with Cursor, Continue.dev, GitHub Copilot, Gemini, Codeium, Tabnine, Cline and any assistant that reads project instruction files.

What you lose without Claude Code: native integration with the 34 Agent Skills (in other assistants the .md files remain accessible but don't auto-load) and the spec-guard.sh hook that ties into the Claude Code runtime (the equivalent protection is covered by the independent pre-push git hook).

What you DON'T lose: CONTRACT, agents, commands, memory-bank, specs, Asana/CodeRabbit integration, the survey, the openspec CLI — all the same.

05 · Switching provider

What if I want to switch AI provider mid-project?

You do it without losing anything. But the "how" depends on your setup, because AI provider, assistant and IDE are three distinct layers:

AI Provider

Anthropic, OpenAI, Google, Meta

Assistant

Claude Code, Continue.dev, Copilot, Cmd-K

IDE

VSCode, Cursor, JetBrains, Vim

If you use Cursor or Continue.dev, switching provider is a click in their settings. If you use Claude Code and want to move to OpenAI, you swap extensions (not necessarily IDEs). If you switch IDE entirely, you open the repo in the new IDE and the full context is still there. The .md and .sh files stay intact no matter what.

06 · Asana

Do I need Asana mandatorily?

For some commands yes (crear-tickets, planificar, status updates inside implementar). For the bulk of the framework, no.

Without Asana these still work: survey, CONTRACT, agents, commands refinar / documentar / explicar / compliance / adoptar, the 34 Agent Skills, the full openspec CLI, memory-bank, global standards and the 3-layer validation pipeline. If your team uses Linear, Jira, Notion or GitHub Issues, you can manage the backlog manually.

07 · Team + CONTRACT

How does it scale with a team? Conflicts on CONTRACT?

The CONTRACT is immutable by design, so classic conflicts on that file don't exist. Changes to the original scope go to ANEXOS.md (append-only, versioned). Technical decisions go to context/decisions.md with date and author.

Each feature in its branch feature/<project>-<ticket>-<desc>, PR against stage, then PR against main. CodeRabbit reviews against CONTRACT + standards. The 3-layer pipeline validates before merge.

08 · TDD / Agile

Does it replace TDD / Agile / Scrum?

No, it complements them. Each operates at a different layer:

  • → SDD (SpecLeap): project's long-term memory — what's being built and why.
  • → TDD: how you write code (test first).
  • → Agile / Scrum: how you organize work over time.

SpecLeap is the missing layer in the methodology stack now that AI writes the code. They coexist without friction: your CONTRACT doesn't clash with your sprint board, your TDD tests don't clash with your specs, and your Scrum Definition of Done is enriched by SpecLeap's 3-layer validation pipeline.

Question not covered? Write to info@specleap.com or open an issue on GitHub.

No excuses

Start
today

100% free. No registration. No credit card. Just clone and start.

View on GitHub → Read Documentation
Community

Collaborate with
SpecLeap

SpecLeap is open source and maintained by the community. Contribute code, docs, or a donation.

Form

Join as a contributor

Share your details and we'll get in touch to onboard you. Code, docs, testing, design: any help counts.