Awesome-omni-skill warping
Apply warping framework standards for AI-assisted development. Use when starting projects, writing code, running tests, making commits, or when the user references warping, project standards, or coding guidelines.
git clone https://github.com/diegosouzapw/awesome-omni-skill
T=$(mktemp -d) && git clone --depth=1 https://github.com/diegosouzapw/awesome-omni-skill "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/development/warping" ~/.claude/skills/diegosouzapw-awesome-omni-skill-warping && rm -rf "$T"
skills/development/warping/SKILL.mdWarping Framework
A layered framework for AI-assisted development with consistent standards and workflows.
When This Skill Activates
This skill automatically loads when you:
- Start work in a warping-enabled project (has
directory)./warping/ - Reference warping, project standards, or coding conventions
- Run tests, make commits, or perform quality checks
- Ask about project structure, workflows, or best practices
Core Principle: Rule Precedence
Warping uses hierarchical rules where more specific overrides general:
user.md ← HIGHEST precedence (personal preferences) ↓ project.md ← Project-specific rules ↓ {language}.md ← Language standards (python.md, go.md, typescript.md, cpp.md) ↓ {tool}.md ← Tool guidelines (taskfile.md, git.md) ↓ main.md ← General AI behavior ↓ specification.md ← LOWEST precedence (requirements)
IMPORTANT: If
user.md says one thing and python.md says another, user.md ALWAYS wins.
File Reading Strategy (Lazy Loading)
DO NOT read all warping files at once. Read only what you need:
- Always start with:
(general guidelines)./warping/main.md - Check for:
(personal overrides - highest precedence)./warping/core/user.md - Check for:
(project-specific rules)./warping/core/project.md - Then read language-specific only if working with that language:
./warping/languages/python.md./warping/languages/go.md./warping/languages/typescript.md./warping/languages/cpp.md
- Read tool files only when using that tool:
(when running tasks)./warping/tools/taskfile.md
(when using git)./warping/scm/git.md
(when using GitHub)./warping/scm/github.md
Task-Centric Workflow
Warping projects use Taskfile as the universal task runner.
Discovery
task --list # See all available tasks task # Same as task --list
Common Tasks
task check # Pre-commit checks (fmt, lint, test, coverage) task test # Run tests task test:coverage # Run with coverage task fmt # Format code task lint # Lint code task build # Build project task clean # Clean artifacts
CRITICAL: Before commits, ALWAYS run
task check. Never claim checks passed without running them.
Test-Driven Development (TDD)
Warping embraces TDD by default:
- Write test first - Define expected behavior
- Watch it fail - Confirm test fails correctly
- Implement - Write minimal code to pass
- Refactor - Improve while keeping tests green
- Repeat - Build incrementally
Coverage Requirements:
- Default: ≥85% coverage (overall + per-module)
- Check
for project-specific thresholdsproject.md - Never claim coverage passes without running
task test:coverage
Spec-Driven Development (SDD)
For new features or projects:
- Build PRD.md - Run
to create Product Requirements Documentwarping.sh spec - AI Interview - Answer focused questions to clarify requirements
- PRD.md Review - Let user review/change PRD.md if they want
- Generate SPECIFICATION.md - use PRD.md + Warping rules to build a complete spec with phases, dependencies, and tasks
- SPECIFICATION.md review -- Let user review/change SPECIFICATION.md if they want
- Implement - Build according to spec, following all applicable warping rules
Quality Standards
Before Every Commit
task check # MUST run this
This typically includes:
- Code formatting
- Linting
- Type checking
- Tests with coverage
- Any project-specific checks
Conventional Commits
ALL commits must follow https://www.conventionalcommits.org/en/v1.0.0/:
feat: add new feature fix: resolve bug docs: update documentation chore: routine tasks test: add/update tests refactor: code restructuring
File Naming
- Use hyphens in filenames, not underscores
- Example:
NOTuser-service.pyuser_service.py - Exception: Language conventions (Python modules can use underscores in code)
Secrets Management
- Store secrets in
directorysecrets/ - Provide
templates.example - Never commit actual secrets
- Use environment variables in code
Language-Specific Standards
Python
- Testing: pytest, ≥85% coverage
- Style: ruff, black, isort (PEP 8)
- Types: mypy strict mode
- Docs: PEP 257 docstrings
Go
- Testing: Testify, ≥85% coverage
- Docs: go.dev/doc/comment
- Patterns: table-driven tests, interface design
TypeScript
- Testing: Vitest/Jest, ≥85% coverage
- Style: ESLint, Prettier
- Types: strict mode, no
any
C++
- Testing: Catch2/GoogleTest, ≥85% coverage
- Standard: C++20/23
- Style: clang-format, clang-tidy
New Project Workflow
When user says "start a new project with warping", "use warping to build X", or similar, follow this complete workflow:
Step 1: Initialize Warping Structure
./warping.sh init
- Creates
directory with framework files./warping/ - Sets up
directorysecrets/ - Creates basic
Taskfile.yml - Adds
for secrets.gitignore
Step 2: User Configuration (First Time Only)
Check if
./warping/core/user.md exists. If not:
./warping.sh bootstrap
- Prompts for user name
- Sets default coverage threshold
- Configures primary languages
- Creates personalized
user.md
Step 3: Project Configuration
./warping.sh project
- Prompts for project name
- Selects project type (CLI, TUI, REST API, Web App, Library, Other)
- Chooses primary language
- Sets coverage threshold
- Creates
with standardsproject.md
Step 4: Specification-Driven Development
./warping.sh spec
This launches the full SDD workflow:
a) Generate PRD.md:
- Prompts for project name, description, and initial features
- Creates
with structured templatePRD.md - Displays full path to PRD.md
b) Conduct AI Interview:
- Read the PRD.md thoroughly
- Ask focused, non-trivial questions to clarify:
- Missing decisions and edge cases
- Implementation details and architecture
- UX considerations and constraints
- Dependencies and tradeoffs
- Each question should have numbered options plus "other"
- Continue until ambiguity is minimized
c) User Review of PRD:
- Let user review/edit PRD.md if they want
- Wait for confirmation before proceeding
d) Generate SPECIFICATION.md:
- Use PRD.md + warping rules to create complete implementation spec
- Include clear phases, subphases, and tasks
- Map dependencies (what blocks what)
- Identify parallel work opportunities
- NO code in specification - just the plan
e) User Review of Specification:
- Let user review/edit SPECIFICATION.md if they want
- Wait for confirmation before implementation
f) Begin Implementation:
- Follow the SPECIFICATION.md plan
- Apply all warping rules (TDD, quality standards, task-centric)
- Maintain ≥85% coverage
- Run
before each committask check - Use Conventional Commits format
Step 5: Development Cycle
- Write tests first (TDD)
- Implement features incrementally
- Run
frequentlytask check - Commit with proper format
- Push changes
Project Initialization (Quick Reference)
For New Projects (Full Workflow)
Use the "New Project Workflow" above when starting from scratch.
For Existing Warping Projects
- Check for
directory./warping/ - Read
for general guidelines./warping/main.md - Read
for user preferences./warping/core/user.md - Read
for project rules./warping/core/project.md - Run
to see available taskstask --list
For New Features in Existing Projects
- Consider running
for complex features./warping.sh spec - Follow SDD process (PRD → Interview → Specification → Implement)
- Or for simple features, just follow TDD directly
Self-Improvement Mechanism
Warping learns and evolves:
- Patterns learned during development (AI can update)meta/lessons.md
- Future improvements noticedmeta/ideas.md
- Project-specific suggestionsmeta/suggestions.md
When you discover a better pattern or make repeated corrections, consider updating
lessons.md.
Safety and Best Practices
Version Control
- Never force-push without explicit permission
- Assume production impact unless stated otherwise
- Prefer small, reversible changes
- Call out risks explicitly
Code Quality
- Run
before EVERY committask check - Verify tests pass:
task test - Check coverage meets threshold:
task test:coverage - Never claim checks passed without running them
Documentation
- Keep docs in
directory, not project rootdocs/ - Update docs when changing behavior
- Use RFC2119 notation in technical docs (!, ~, ?, ⊗, ≉)
Working with Warping
First Time in a Project
- Check for
directory./warping/ - Read
./warping/main.md - Check
for user preferences./warping/core/user.md - Check
for project rules./warping/core/project.md - Run
to see available taskstask --list
During Development
- Write tests first (TDD)
- Implement features
- Run
frequentlytask check - Before commit: ALWAYS
task check - Use Conventional Commits format
- Push changes
When Stuck
- Check
for guidance on which files to read./warping/REFERENCES.md - Review
for learned patternsmeta/lessons.md - Consult language-specific files in
./warping/languages/ - Check
for available commandsTaskfile.yml
Example Workflows
Starting a Feature
# 1. Write test first (TDD) # 2. Watch it fail task test # 3. Implement feature # 4. Run checks task check # 5. Commit git commit -m "feat: add new feature"
Code Review
# 1. Run quality checks task check # 2. Review coverage task test:coverage # 3. Check commit format (Conventional Commits) git log --oneline -n 5 # 4. Suggest improvements → meta/suggestions.md
Bug Fix
# 1. Write failing test that reproduces bug # 2. Fix code # 3. Verify test passes task test # 4. Run full checks task check # 5. Commit git commit -m "fix: resolve issue with X"
Integration Notes
This SKILL.md follows the AgentSkills specification, making it compatible with multiple AI platforms.
With Claude Code
- This SKILL.md file teaches Claude Code about warping
- Place in
(personal) or~/.claude/skills/warping/
(project).claude/skills/warping/ - Claude automatically applies these rules when relevant
- Works in VS Code via Claude Code extension
With clawd.bot
- Compatible with clawd.bot's AgentSkills system
- Place in
(shared) or~/.clawdbot/skills/warping/
(per-agent)<workspace>/skills/warping/ - Install via ClawdHub:
(once published)clawdhub sync warping - Works across WhatsApp, Telegram, Discord, and other channels
- Requires
binary (specified in metadata)task - Supports macOS and Linux (specified in
field)os
With Warp AI
- Upload warping files to Warp Drive
- Create Warp rules referencing warping/*.md files
- Use
orWARP.md
in project rootAGENTS.md
Skill Locations
Claude Code:
- Personal:
~/.claude/skills/warping/SKILL.md - Project:
.claude/skills/warping/SKILL.md
clawd.bot:
- Shared (all agents):
~/.clawdbot/skills/warping/SKILL.md - Per-agent:
<workspace>/skills/warping/SKILL.md - Via registry:
clawdhub sync warping
Warping files:
./warping/*.md (framework files)
Publishing to Registries
Claude Code Skills Marketplace:
- Browse: https://skillsmp.com
- Submit via GitHub repository
- Tag with
,claude-skillsai-coding
ClawdHub (clawd.bot):
- Browse: https://clawdhub.com
- Publish:
clawdhub publish - Update:
clawdhub publish --version x.y.z
Multi-Platform Benefits
- Same format - Write once, use everywhere
- Consistent standards - Warping rules apply across all AI assistants
- Universal workflows - TDD, SDD, quality checks work in any context
- Shared knowledge - Updates propagate to all platforms
Quick Reference
| Task | Command |
|---|---|
| List tasks | or |
| Pre-commit checks | |
| Run tests | |
| Check coverage | |
| Format code | |
| Lint code | |
| Initialize warping | |
| Configure user | |
| Configure project | |
| Generate spec | |
Remember
- Lazy load files - Only read what you need
- User.md is king - Highest precedence always
- Task-centric - Use
for everythingtask - Test first - Write tests before implementation
- Always check - Run
before commitstask check - Conventional commits - Follow the standard
- Coverage matters - ≥85% by default
- Never lie - Don't claim checks passed without running them
For more details, read the specific files in
./warping/ as needed. Start with main.md and follow the precedence hierarchy.