[opensourcedex]

claude-code-skills is a free, open source version control & collaboration project written in Shell and released under MIT. It has 587 GitHub stars, 88 forks and 0 open issues, and was last pushed 7 days ago. On this registry it ranks #44 of 50 tracked projects in Version Control & Collaboration, with 5 head-to-head comparisons available. It gained 21 stars over the last 30 tracked days.

What is claude-code-skills?

Claude-Codex Skills is an MIT-licensed collection of installable agent skills for Claude Code and Codex that gives an AI agent a clear finish line: establish the intended outcome, work within scope, verify the result, and make remaining gaps explicit.

What it is

Claude-Codex Skills is a registry of standalone skills, packaged as plugins and distributed through a plugin marketplace, that live inside the Claude Code and Codex ecosystems. Each skill addresses one stage of software work, running from deciding whether an idea is worth building through architecture, delivery planning, implementation, quality assurance, delivery, and operations, with a separate family that maintains the collection itself. The project is written primarily in PowerShell and carries the topics agent-skills, ai-agents, and ai-driven-development alongside claude-code, codex, code-review, and codebase-audit. The registry lists it under Developer Tools / Version Control & Collaboration, with 565 stars and 84 forks.

The concrete problem it solves is the gap between the request and the finished work. A user asks for a fix and receives a new abstraction; a review returns generic advice; the agent declares the task done without saying what it checked. These plugins replace the loose, one-off instruction typed into a chat with a named plugin installed from a marketplace and invoked by a full skill name, so the agent works against a stated target, a bounded scope, and a report of what was verified. Implementation carries its own verification, which makes a separate test-building or review skill optional rather than mandatory.

Key capabilities

  • Eight plugin families cover the lifecycle: product-discovery-suite, architecture-suite, delivery-planning-suite, implementation-suite, quality-assurance-suite, delivery-suite, operations-suite, and skill-maintenance-suite.
  • Installation runs through marketplace commands: /plugin marketplace add levnikolaevich/claude-code-skills, then /plugin install implementation-suite@levnikolaevich-skills-marketplace, then /reload-plugins.
  • The Codex path uses codex plugin marketplace add levnikolaevich/claude-code-skills and codex plugin add implementation-suite@levnikolaevich-skills-marketplace.
  • Skills are invoked by full namespaced name, such as /implementation-suite:ln-41-surgical-change-implementer in Claude Code or $ln-41-surgical-change-implementer in Codex.
  • Skills work independently, so a small fix does not require a full lifecycle or audit.
  • Consequential actions stay inside authorization boundaries: reviews report findings, implementation stays within the agreed scope, and publication and deployment follow their own boundaries.

Who uses it and how

  • A solo developer or small team with a bounded fix installs implementation-suite alone, skipping the full lifecycle.
  • A larger task combines only the steps needed to resolve decisions and produce evidence.
  • Teams starting from supplied requirements, existing designs, or current tests treat those as valid starting points rather than restarting discovery.
  • Release owners use delivery-suite to update an authorized destination and check the observed result.
  • Maintainers of the collection use skill-maintenance-suite to review boundaries, consistency, and distribution readiness.

Getting started

Install from the marketplace, beginning with implementation-suite for a feature or fix: /plugin marketplace add levnikolaevich/claude-code-skills followed by /plugin install implementation-suite@levnikolaevich-skills-marketplace and /reload-plugins in Claude Code. For Codex, use codex plugin marketplace add and codex plugin add, and replace implementation-suite with another plugin from the route table for a different task.

How it compares

The provided facts name no paid products this project replaces, and they name no comparable tools, so Claude-Codex Skills stands alone in this registry under Developer Tools / Version Control & Collaboration. Its distinguishing position is a single set of skills, installed from one marketplace, that targets both Claude Code and Codex.

When to use it — and when not to

It is a skills collection rather than a runtime, so anyone who does not work inside Claude Code or Codex gains nothing from installing it, and the user must still supply the intended outcome for the agent to check against. The repository is written primarily in PowerShell and carries a permissive MIT licence with no open issues at the last push, so there is no commercial support tier or vendor SLA behind it — teams that require contractual support or a vendor-backed roadmap should look elsewhere.

project readme (upstream, from github) — read inline

Claude-Codex Skills

Give your AI agent a clear finish line.

You ask for a fix and get a new abstraction. A review lists generic advice. The agent says “done,” but you still have to work out what it checked.

These skills give Claude Code and Codex a concrete way to finish the task: establish the intended outcome, work within scope, verify the result, and make remaining gaps explicit. Choose the skill for the problem in front of you.

Project page · Install · Full catalog

What changes in your workflow

  • A clearer target. Turn an idea, bug report, or incident into an outcome the agent can check.
  • Focused work. Reuse existing capabilities, fix the owning cause, and remove what the change makes obsolete.
  • A result you can assess. See the evidence, failed or unavailable checks, and remaining risks behind the verdict.
  • Control over consequential actions. Reviews report findings; implementation stays within the agreed scope; publication and deployment follow their authorization boundaries.

Choose a route

Start where you need help. Each skill works independently; a small fix does not require a full lifecycle or audit.

Your problem Plugin to install Useful result
Is this idea worth building? product-discovery-suite A reasoned opportunity decision, clear requirements, or a usable interaction design.
How should this system change? architecture-suite An evidence-backed system map, design decision, or migration approach.
Is this plan ready to implement? delivery-planning-suite Missing decisions exposed, work sequenced, and verification planned.
Fix, upgrade, simplify, or speed up this code. implementation-suite A bounded change checked against the intended behavior and benefit.
What could go wrong with this delivery or codebase? quality-assurance-suite Prioritized findings tied to evidence, or acceptance tests for the agreed scope.
Get this change published or deployed. delivery-suite The authorized destination updated and the observed result checked.
Why did this fail, or did the feature help? operations-suite An evidence-backed diagnosis or product outcome assessment.
Can I trust the instructions in this skill? skill-maintenance-suite A review of its boundaries, consistency, and distribution readiness.

For a larger task, combine only the steps needed to resolve decisions and produce evidence. Supplied requirements, existing designs, and current tests are valid starting points. Implementation includes its own verification; a separate test-building or review skill is optional.

Install

For a feature or fix, start with implementation-suite. Replace it with another plugin from Choose a route for a different task.

Claude Code:

/plugin marketplace add levnikolaevich/claude-code-skills
/plugin install implementation-suite@levnikolaevich-skills-marketplace
/reload-plugins

Codex:

codex plugin marketplace add levnikolaevich/claude-code-skills
codex plugin add implementation-suite@levnikolaevich-skills-marketplace

Invoke a skill by its full name: /implementation-suite:ln-41-surgical-change-implementer in Claude Code or $ln-41-surgical-change-implementer in Codex, followed by your task. For example: “Fix the total calculation when an item has a discount. Preserve the current rounding rules.”

Lifecycle and plugins

Standalone skills for Claude Code and Codex, grouped by the work you need to do. Browse the outcomes below to find the right skill for your task.

1 Product discovery → 2 Architecture → 3 Delivery planning
 → 4 Implementation → 5 Quality assurance → 6 Delivery → 7 Operations
 ↑ Product feedback / architecture changes / operational remediation ↵

8 Skill maintenance supports the collection itself.

The numbers group skills by task. Install one family or combine the capabilities you need.

Product Discovery Suite

Decide what is worth building, who it serves, and what a useful first version must do.

Index Skill Purpose
11 Opportunity Evaluator Evaluates new product opportunities through demand, channels and economics before committing to build.
12 Product Requirements Builder Defines product requirements, business rules and acceptance criteria for a committed intent; edits product docs only.
13 Interaction Design Builder Designs user flows, interaction states and mockups for a defined product scope; does not implement UI code.

Architecture Suite

Understand the system you have and choose changes with clear boundaries, tradeoffs, and recovery paths.

Index Skill Purpose
21 System Design Baseline Builder Defines measurable architecture drivers and constraints before system design; edits architecture docs only.
22 Current Architecture Documenter Documents current architecture from implementation evidence; does not propose a target or audit fitness.
23 System Design Proposal Builder Designs target system boundaries, contracts and tradeoffs from requirements; does not plan tasks or implement.
24 Architecture Decision Recorder Records one architecture decision with alternatives, consequences and status; does not design the whole system.
25 Architecture Diagram Builder Creates evidence-backed current or target architecture diagrams; not UI design.
26 Architecture Migration Planner Plans architecture migrations with compatibility, data safety, rollout and recovery; does not execute them.

Delivery Planning Suite

Find missing decisions and risky dependencies before they become implementation rework.

Index Skill Purpose
31 Delivery Plan Builder Builds dependency-ordered delivery plans from requirements and repository evidence; read-only.
32 Test Strategy Planner Plans risk-based test portfolios and acceptance evidence; does not write or execute tests.
33 Plan Reviewer Reviews a concrete implementation plan for missing decisions, feasibility and risk before execution; read-only.

Implementation Suite

Fix the cause, reuse what already works, and keep changes focused on a verified outcome.

Index Skill Purpose
41 Surgical Change Implementer Implements one scoped feature or fix through the smallest complete solution; not upgrades or performance tuning.
42 Dependency Upgrader Upgrades dependencies in reversible batches with version-specific compatibility and regression checks.
43 Code Modernizer Modernizes a bounded capability to reduce demonstrated maintenance cost; not routine upgrades or performance tuning.
44 Performance Optimizer Profiles and improves a measured performance bottleneck; retains only verified improvements.
45 Benchmark Comparator Compares tools or implementations through controlled benchmarks and independent correctness checks.

Quality Assurance Suite

Find gaps that could break a release, mislead a maintainer, or make a passing test meaningless.

Index Skill Purpose
51 Acceptance Test Builder Builds, updates or retires scoped acceptance tests and verifies execution; does not repair product code.
52 Delivery Reviewer Reviews a completed change for acceptance, regressions and release risk; read-only, not a whole-codebase audit.
53 Documentation Auditor Audits documentation and comments for trust, coverage, consistency and freshness; read-only.
54 Codebase Auditor Audits cross-cutting codebase health, security and maintainability; not a single-change review or specialist audit.
55 Test Suite Auditor Audits existing tests for risk coverage, reliable oracles and maintenance value; does not edit tests.
56 Architecture Auditor Audits implemented architecture boundaries, dependencies and ownership; not target design or plan review.
57 Persistence Auditor Audits queries, transactions, consistency and persistence resource lifetimes; read-only.

Delivery Suite

Publish the intended changes and verify what reached the remote repository or target environment.

Index Skill Purpose
61 Repository Publisher Commits, pushes and remotely verifies authorized repository changes; does not create releases.
62 Release Publisher Prepares and publishes an explicitly requested tagged GitHub release; does not deploy applications.
63 Deployment Engineer Prepares CI/CD and infrastructure, then executes authorized deployments with health and recovery checks.
64 Community Announcer Drafts or publishes authorized, fact-checked GitHub Discussions announcements; does not create releases.

Operations Suite

Turn incidents and product signals into evidence for the next recovery or product decision.

Index Skill Purpose
71 Operations Investigator Diagnoses incidents from operational evidence and proposes recovery; does not change live systems.
72 Product Outcome Evaluator Evaluates observed product outcomes against a prior hypothesis; does not run experiments or change user treatment.

Skill Maintenance Suite

Make skills easier to invoke correctly, follow consistently, and verify before distribution.

Index Skill Purpose
81 Skill Reviewer Reviews skill instructions, trigger boundaries and distribution contracts; not product code.

License

MIT

Frequently asked questions

Is claude-code-skills free to use?

claude-code-skills is open source under the MIT licence. There is no licence fee and no seat count — you can self-host it or, where the project offers one, pay a vendor for a managed version instead.

What does claude-code-skills do?

Help your AI agent finish the job: solve the right problem, keep changes focused, and show what was verified. For Claude Code and Codex.

What is claude-code-skills written in?

claude-code-skills is primarily written in Shell. Its source is publicly available at https://github.com/levnikolaevich/claude-code-skills, and it has 587 GitHub stars.