DEDUPED Hot AI repos. Real use cases. What people build. Get the Monday 5

BEST AI REPOS BY TASK

Best open-source AI repos for skills and workflows for coding agents (2026)

8 repos compared · updated 2026-10-27

What are the best skills and plugins for Claude Code, Codex and other coding agents?

The best open-source AI repos for skills and workflows for coding agents right now are Superpowers, Matt Pocock's Skills and Spec Kit. Superpowers and Spec Kit give your agent a full method (design, plan, build, test); Matt Pocock's and Addy Osmani's skill sets are smaller, editable building blocks you can adopt one at a time.

Comparison

RepoWhat it doesStarsStart here
Superpowers
obra/superpowers · Shell		Superpowers is a set of composable skills and instructions that make coding agents follow a structured software development method: design first, then a written plan, test-driven implementation, review, and finishing the branch.300,763Build a feature from design to reviewed code Intermediate, 1 hour
Matt Pocock's Skills
mattpocock/skills · Shell		A set of editable agent skills for planning, aligning, building, and debugging software with coding agents, taken from the author's .agents directory.295,582Align an agent on a change before any code is written Intermediate, 30 minutes
Spec Kit
github/spec-kit · Python		Spec Kit is GitHub's open-source toolkit that gives AI coding agents structured processes: Spec-Driven Development (constitution, specify, plan, tasks, implement, converge), a bug fixing extension (assess, fix, test) and an idea assessment extension that ends in a go, clarify or stop decision.142,760Build a feature from a spec with your coding agent Intermediate, 1 hour
Agent Skills (Addy Osmani)
addyosmani/agent-skills · JavaScript		Agent Skills is a set of 25 production-grade engineering skills for AI coding agents by Addy Osmani.110,782Go from idea to tested feature with /spec, /plan and /build Intermediate, 1 hour
ECC
affaan-m/ECC · JavaScript		ECC is an MIT-licensed agent harness of skills, agents, hooks, memory, and AgentShield security scanning that runs with Claude Code, Codex, and other coding agents.285,237Install ECC as a Claude Code plugin in one guided run Beginner, 15 minutes
Ponytail
DietrichGebert/ponytail · JavaScript		Ponytail is a prompt and skill set that makes AI coding agents write less code and review changes for bugs, security, and missing tests.172,653Review agent-written changes before merging Intermediate, 30 minutes
wshobson Agents
wshobson/agents · Python		wshobson/agents is a catalog of 94 plugins with 202 agents, 185 skills and 105 commands that installs across Claude Code, Codex CLI, Cursor, OpenCode, Antigravity, GitHub Copilot and Pi from one shared Markdown source.41,537Add the Python development plugin to Claude Code Beginner, 15 minutes
Caveman
JuliusBrussee/caveman · Go		Caveman is a skill and proxy for coding agents that cuts tokens by making the agent talk tersely and read less, while code stays exact.114,855Cut your coding agent's token use with one skill Beginner, 5 minutes

When to pick which

Superpowers obra/superpowers

Superpowers is a set of composable skills and instructions that make coding agents follow a structured software development method: design first, then a written plan, test-driven implementation, review, and finishing the branch.

Best for: Software developers who use a coding agent such as Claude Code, Codex, Cursor, or Gemini CLI and want its work to follow a repeatable design, plan, test, and review process.

Needs: A supported coding agent, such as Claude Code, Codex CLI, Cursor, or Gemini CLI, with Superpowers installed for that agent; A git project to work in; A feature or change you can describe in a few sentences

Start here: Build a feature from design to reviewed code Intermediate, 1 hour, works with claude-code, codex, cursor, gemini-cli

  1. Install Superpowers in your agent. In Claude Code, run /plugin install superpowers@claude-plugins-official.
  2. Describe the feature to the agent and answer its questions so brainstorming can draw out what you want.
  3. Read the design in the short chunks the agent shows you, then sign off on it.
  4. Review the implementation plan from writing-plans, checking that each task lists exact file paths and verification steps.

All 4 use casesGitHub

Matt Pocock's Skills mattpocock/skills

A set of editable agent skills for planning, aligning, building, and debugging software with coding agents, taken from the author's .agents directory.

Best for: Working software engineers who use Claude Code, Codex, or other coding agents and want more control over planning, testing, and design than full process frameworks give.

Needs: A coding agent that supports the plugin or skills install route, such as Claude Code or Codex; A project repository the agent can read and edit; Optional: an issue tracker such as GitHub or GitLab, or a local folder for tickets

Start here: Align an agent on a change before any code is written Intermediate, 30 minutes, works with claude-code, codex

  1. Install the plugin in Claude Code with claude plugin install mattpocock-skills@claude-plugins-official.
  2. In your repo, run /setup-matt-pocock-skills once and answer the questions about issue tracker, triage labels, and where docs go.
  3. Run /grill-with-docs and describe the change you want.
  4. Answer the agent's questions until the plan is clear, and let it update the glossary and ADRs as you go.

All 5 use casesGitHub

Spec Kit github/spec-kit

Spec Kit is GitHub's open-source toolkit that gives AI coding agents structured processes: Spec-Driven Development (constitution, specify, plan, tasks, implement, converge), a bug fixing extension (assess, fix, test) and an idea assessment extension that ends in a go, clarify or stop decision. You install the specify CLI, pick your coding agent integration, then drive the process with slash commands.

Best for: Developers and teams using Copilot, Claude Code or other coding agents who want plans and specs instead of one-shot prompts.

Needs: uv to install the specify CLI; A supported coding agent (for example GitHub Copilot)

Start here: Build a feature from a spec with your coding agent Intermediate, 1 hour, works with claude-code, codex, cursor, gemini-cli

  1. Install the CLI: uv tool install specify-cli
  2. Create a project: specify init my-project --integration copilot
  3. Set principles with /speckit-constitution and describe the feature with /speckit-specify
  4. Pick the stack with /speckit-plan, then run /speckit-tasks and /speckit-implement

All 3 use casesGitHub

Agent Skills (Addy Osmani) addyosmani/agent-skills

Agent Skills is a set of 25 production-grade engineering skills for AI coding agents by Addy Osmani. Nine slash commands (/spec, /plan, /build, /test, /constraints, /review, /webperf, /code-simplify and /ship) map to the development lifecycle and activate the right skills, such as test-driven development and five-axis code review, so agents follow senior engineering workflows consistently.

Best for: Developers using Claude Code, Cursor, Codex, Gemini CLI or Copilot who want their agent to spec, test and review like a senior engineer.

Needs: Node.js for npx skills, or Claude Code, Codex, Gemini CLI or Antigravity for their native installs; Git; SSH keys for the Claude Code marketplace, or the HTTPS URL workaround in the README

Start here: Go from idea to tested feature with /spec, /plan and /build Intermediate, 1 hour, works with claude-code, cursor, codex

  1. Run npx skills add addyosmani/agent-skills to install all 25 skills into your agent
  2. Run /spec and describe what to build; the spec comes before any code
  3. Run /plan to break it into small, atomic tasks
  4. Run /build to implement one slice at a time, or /build auto to implement every task after one plan approval

All 3 use casesGitHub

ECC affaan-m/ECC

ECC is an MIT-licensed agent harness of skills, agents, hooks, memory, and AgentShield security scanning that runs with Claude Code, Codex, and other coding agents.

Best for: Developers and engineering teams who already use Claude Code or Codex and want a repeatable plan, test, review, and remember workflow installed once.

Needs: Node.js 18 or newer; Git on PATH; Claude Code 2.1 or newer for plugin setup; Access to github.com and npm to fetch the official ECC packages

Start here: Install ECC as a Claude Code plugin in one guided run Beginner, 15 minutes, works with claude-code

  1. Install Node.js 18 or newer, Git, and Claude Code, then confirm with node --version, git --version, and claude --version.
  2. Run npx [email protected] setup.
  3. Choose Global user scope and Standard hooks for a typical personal setup, then confirm.
  4. Start a new Claude Code session and run /plugin list to check that ecc@ecc is enabled.

All 4 use casesGitHub

Ponytail DietrichGebert/ponytail

Ponytail is a prompt and skill set that makes AI coding agents write less code and review changes for bugs, security, and missing tests.

Best for: Developers and engineering leads who use AI coding agents and want smaller, reviewed changes.

Needs: Claude Code or Codex installed; A git repository with a project to work on; Permission to install a plugin into your agent

Start here: Review agent-written changes before merging Intermediate, 30 minutes, works with claude-code, codex

  1. Run /plugin marketplace add DietrichGebert/ponytail in Claude Code.
  2. Run /plugin install ponytail@ponytail in Claude Code.
  3. Let the agent make your change in the repo.
  4. Run /ponytail-review uncommitted to review the current uncommitted changes.

All 4 use casesGitHub

wshobson Agents wshobson/agents

wshobson/agents is a catalog of 94 plugins with 202 agents, 185 skills and 105 commands that installs across Claude Code, Codex CLI, Cursor, OpenCode, Antigravity, GitHub Copilot and Pi from one shared Markdown source.

Best for: Developers who want ready-made agents, skills and commands for Python, JavaScript, testing, infrastructure and security work in their coding tool.

Needs: A supported harness such as Claude Code or Codex CLI; GitHub CLI or Node for single-skill installs

Start here: Add the Python development plugin to Claude Code Beginner, 15 minutes, works with claude-code

  1. In Claude Code run /plugin marketplace add wshobson/agents
  2. Install the plugin: /plugin install python-development@claude-code-workflows
  3. Scaffold a service: /python-development:python-scaffold Create a FastAPI service with tests
  4. Read the generated layout and dependencies before running anything

All 3 use casesGitHub

Caveman JuliusBrussee/caveman

Caveman is a skill and proxy for coding agents that cuts tokens by making the agent talk tersely and read less, while code stays exact. The README reports 33.2% fewer input tokens across 54 Claude Code runs and links an Adobe Research paper on caveman-style output; it also links a page listing every case where caveman loses. It installs as a skill, a Claude Code plugin, a Gemini CLI extension or an all-agents installer, with an optional free Auto model router.

Best for: Developers who pay per token for Claude Code, Codex or Gemini CLI and want shorter answers and smaller context without losing correctness.

Needs: A coding agent such as Claude Code, Codex or Gemini CLI; Node.js for the npx installs (22.13+ for the all-agents installer)

Start here: Cut your coding agent's token use with one skill Beginner, 5 minutes, works with claude-code, codex, gemini-cli

  1. Install for every agent with npx skills add JuliusBrussee/caveman -g
  2. Or in Claude Code: /plugin marketplace add JuliusBrussee/caveman then /plugin install caveman@caveman
  3. Type /caveman ($caveman in Codex) if it does not start on its own; /ultracave is even shorter
  4. Say stop caveman to go back to normal

All 3 use casesGitHub

Use cases for skills and workflows for coding agents

Questions

What are the best skills and plugins for Claude Code, Codex and other coding agents?

The best open-source AI repos for skills and workflows for coding agents right now are Superpowers, Matt Pocock's Skills and Spec Kit. Superpowers and Spec Kit give your agent a full method (design, plan, build, test); Matt Pocock's and Addy Osmani's skill sets are smaller, editable building blocks you can adopt one at a time.

Which repo for skills and workflows for coding agents is easiest to start with?

ECC: its starter use case "Install ECC as a Claude Code plugin in one guided run" is rated Beginner and takes about 15 minutes.

How do I get started with Superpowers?

Install Superpowers in your agent. In Claude Code, run /plugin install superpowers@claude-plugins-official; Describe the feature to the agent and answer its questions so brainstorming can draw out what you want; Read the design in the short chunks the agent shows you, then sign off on it. Full steps and a copy-paste agent prompt: https://dedupedai.com/use-cases/obra-superpowers/build-feature-from-design-to-review/

Are these skills and workflows for coding agents repos free?

They are open-source projects on GitHub (check each license; some, like fair-code projects, limit commercial use). Some also sell a hosted version or need a paid model API key; the requirements for each repo are listed on this page.

How is this list of skills and workflows for coding agents repos chosen?

Deduped tracks the hottest AI repos on GitHub and writes step-by-step use cases from each README. For this page we picked the repos that do this job, ordered by fit and then GitHub stars. Repos we only inferred from a short README, or that we do not index, are left out. The page is rebuilt with the site; stars are as of the last build in 2026.

More answers

For AI agents: Markdown version · JSON · MCP server · llms.txt

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer