Open-source project
daymade/claude-code-skills avatar
daymade/claude-code-skills

daymade/claude-code-skills: A Marketplace for Claude Code Skills with a Hardened skill-creator

Professional Claude Code skills marketplace featuring production-ready skills for enhanced development workflows.

1,392 stars219 forksPythonMIT

At a glance

What is it?
This repository packages production-ready Claude Code skills as plugins, with a focus on a fork of Anthropic's skill-creator that adds validation, security gates, and concurrency safeguards. The core judgement: it is useful for teams building custom skills, but its value depends on the maturity of the individual skills.
Who is it for?
Adopt this marketplace if you are building custom Claude Code skills and need guardrails beyond Anthropic's official skill-creator: the fork adds security scans, validation, and concurrency handling that the official version lacks. Do not adopt it if you only want ready-made skills for common tasks, because the quality of each skill varies and the README does not document them in depth.
Can I use it commercially?
Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
Is it still maintained?
Yes. The repository received new commits within the last day.
What is it written in?
Mainly Python, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on September 15, 2026, and from our analysis. They are not legal advice.

DEEP OPEN-SOURCE ANALYSIS

What This Marketplace Solves

The repository addresses a specific gap: turning Claude Code skills from ad-hoc prompts into distributable, validated plugins. The README positions it as a professional marketplace, and the primary artifact is the skill-creator, a fork of Anthropic's official skill-creator. The official version tells you what to build, while this fork claims to tell you what not to try and why. The target user is a developer or team that wants to create, package, and maintain their own skills without hitting the undocumented failure modes that come from real-world usage. The repository also bundles other skills, like a usage analyst and a bilibili source, but the skill-creator is the centerpiece and the reason to install the marketplace.

The skill-creator: A Fork with Guardrails

The skill-creator here is not a cosmetic change. The README lists specific additions over the official version. It includes an ordered search protocol with a decision matrix that forces you to choose Adopt, Extend, or Build before writing anything. It uses structured AskUserQuestion checkpoints instead of prose instructions, which the README says keeps the user from losing context. The fork also adds a security scan gate via security_scan.py with gitleaks integration, which runs before packaging. That is a hard gate, not a suggestion. The validation step is expanded beyond a basic YAML check: it includes a provenance-checked capability audit that compares old and new versions, and the packaging step re-verifies the completed review instead of trusting a marker. This is a deliberate attempt to catch mistakes that the official tool would let through.

Concurrency and Distillation: Handling Real Repo Work

Two areas stand out as genuinely different from the official skill-creator. First, concurrent sessions on the same skill repo. The fork uses git-ref baselines, re-read-before-write, a HEAD check before commit, and own-paths-only staging. That means if you run multiple Claude Code sessions against one skill directory, the tool tries to prevent clobbering. Second, the artifact-corpus distillation workflow. This is for turning past conversations into skill behavior. The README describes a script-measured pattern extraction with a threshold of at least three artifacts, a gate that distinguishes cataloging from distillation, and an independent completeness audit. It also imposes evidence-boundary rules, like an authority ladder that ranks real calls and machine-readable specs above production code. These are not features you would guess from the description; they are responses to specific failures the maintainer apparently hit.

Installation: Plugins, Not Just Copy-Paste

Installation is via Claude Code's plugin system, not a simple git clone. The in-app command is `/plugin marketplace add daymade/claude-code-skills`, then you browse and install the skill-creator plugin. From the terminal, you run `claude plugin marketplace add https://github.com/daymade/claude-code-skills` and then `claude plugin install daymade-skill@daymade-skills`. The README is explicit about a common mistake: the marketplace name is `daymade-skills` from marketplace.json, and using the repo path as a marketplace name, like `@daymade/claude-code-skills`, will fail. There is also an automated installer via a curl script for macOS/Linux and a PowerShell script for Windows. The documentation suite installs under a shared namespace, `daymade-docs`, which exposes skills like `doc-to-mark`. The plugin model means each skill is a separate install, so you do not have to load the entire marketplace into every session.

A Genuine Limitation: The Marketplace Is Only as Good as Each Skill

The README is long on the skill-creator and short on the other skills. The recent releases mention a claude-usage-analyst and a bilibili-source, and there is a pdf-creator with CJK coverage, but the README does not describe their capabilities, failure modes, or maintenance status. That is a real limitation. You are adopting a marketplace, but the only skill with deep documentation is the meta-skill. The other skills could be thin wrappers or abandoned. The concurrency handling in skill-creator suggests the maintainer cares about robustness, but that does not extend automatically to the rest of the collection. Also, the security scan is a gate for skills you create, not for skills you install. Nothing in the README says the bundled skills have passed the same gitleaks scan. If you install a third-party skill from this marketplace, you are trusting the maintainer's review process, which is not documented for non-skill-creator skills.

The Alternative: Anthropic's Official skill-creator

The direct alternative is the official skill-creator from Anthropic's skills repository. The README frames this fork as production-hardened, and the comparison table shows the differences. The official version tells you what to build but gives no guidance on what not to try. It does not have a security scan gate, no concurrency handling, and no artifact-corpus distillation workflow. The official version also does not handle the clash detection when it coexists with this fork's skill-creator. That clash is a real problem: the two descriptions are near-identical, and the fork detects the clash on trigger and offers a SessionStart routing hook. If you are a solo developer creating a simple skill, the official version may be enough and is simpler. If you are working on a team with multiple sessions or need to package skills for distribution, the fork's guardrails save time. The trade-off is complexity: the fork's methodology document is a reference, not a quick read.

Maintenance and License Implications

The repository is under the MIT license, which means you can fork and modify it freely, including the skill-creator, with attribution. The last push was June 2026, and releases are frequent, with v1.64.0 on June 13, 2026, and v1.60.1 on June 4, 2026. That suggests active maintenance. The release notes show a pattern of hardening, such as 'Harden macos-cleaner safety' in v1.60.1, which indicates the maintainer responds to issues. However, the README does not state a deprecation policy for skills. If a skill becomes obsolete, there is no documented process for removal or migration. The upgrade cost is tied to the plugin system: you update via the same plugin commands, but the README does not explain what happens to your existing skills when a new version ships. For a skill-creator that you use to build other skills, a breaking change in the validation rules could force you to re-validate all existing skills. That is a cost to plan for, but the git-ref baselines and versioning rules in the fork suggest the maintainer is aware of upgrade friction.

Editorial conclusion

Adopt this marketplace if you are building custom Claude Code skills and need guardrails beyond Anthropic's official skill-creator: the fork adds security scans, validation, and concurrency handling that the official version lacks. Do not adopt it if you only want ready-made skills for common tasks, because the quality of each skill varies and the README does not document them in depth. Before installing, verify the specific skill you need exists in the repository, check its SKILL.md for maintenance status, and confirm the plugin installation path matches the documented `@daymade-skills` name, as the README warns that using the repo path as a marketplace name fails.

Official sources

  1. Official README
  2. Project repository
  3. Release notes
Community notes

Community notes