Skip to content
/

daymade/claude-code-skills

daymade-cc-skills · daymade/claude-code-skills · ★ 1.1k · last commit 2026-05-24

Primitive shape 52 total
Skills 52
00

Summary

daymade/claude-code-skills — Summary

daymade/claude-code-skills is a 52-skill Claude Code marketplace focused on production-hardened, battle-tested skill implementations, many of which are forks of official Anthropic skills with significant improvements. The flagship skill-creator is described as a "production-hardened fork of Anthropic's official skill-creator" with 8-channel research protocol, 9 structured AskUserQuestion checkpoints, cache edit warnings, security scan gate (gitleaks), and expanded validator. A skill-reviewer provides three-mode review (self-review, external review, auto-PR). A skills-search skill discovers existing skills before building new ones. Most skills include Python scripts for automation (init_qa_project.py, security_scan.py, quick_validate.py) and references/ subdirectories with detailed methodology documents. The qa-expert skill implements Google Testing Standards + OWASP with autonomous execution via a "master QA prompt." The repo is China-bilingual (README.zh-CN.md) and includes a Claude Code-centric vertical (daymade-claude-code plugin with 7 Claude-Code-specific skills). Compared to seeds, daymade sits closest to spec-kit (quality-over-quantity with methodological depth) but focuses on skill craftsmanship rather than spec-driven workflow, and uniquely focuses on improving the skill-creation and review process itself.

01

Overview

daymade/claude-code-skills — Overview

Origin

Created by daymade. 1,111 stars, last push 2026-05-24. v1.52.0. Bilingual (English + Chinese), indicating Chinese developer community focus.

Philosophy

From README:

"The official skill-creator tells you what to build. Ours also tells you what not to try — and why."

Quality comparison table (independent audit, 8 dimensions):

Dimension Official This Fork
Actionability 7 9
Error Prevention 5 9
Prior Art Research 4 9
Counter Review Process 4 8
Real-World Lessons 3 8
User Experience 4 9
Total (out of 80) 42 65

Key philosophical stance: quality > quantity. 52 skills vs thousands in other repos, but each skill is heavily documented with references, scripts, evaluation checklists, and security gates.

Skill-Creation Meta-Skills

Three meta-skills for the skill development process:

  1. skill-creator — create skills with 8-channel research + 9 structured checkpoints
  2. skill-reviewer — review/improve skills (self, external, or auto-PR modes)
  3. skills-search — discover existing skills before building new ones

Claude Code Focus

The daymade-claude-code plugin is a vertical of 7 skills specific to Claude Code usage:

  • claude-code-history-files-finder — find historical context files
  • claude-export-txt-better — improve Claude conversation exports
  • claude-md-progressive-disclosurer — implement progressive disclosure in CLAUDE.md
  • continue-claude-work — resume interrupted work sessions
  • statusline-generator — generate status line displays
  • claude-skills-troubleshooting — debug skill issues
  • marketplace-dev — marketplace development tools
02

Architecture

daymade/claude-code-skills — Architecture

Distribution

Claude Code plugin marketplace.

/plugin marketplace add daymade/claude-code-skills

# Install individual skills
claude plugin install skill-creator@daymade-skills
claude plugin install qa-expert@daymade-skills

Directory Structure

<skill-name>/          # One directory per skill (45+ plugin dirs)
  SKILL.md             # Skill instructions
  scripts/             # Python automation scripts
  references/          # Deep-dive reference docs
  assets/              # Templates

daymade-skill/         # Meta-skill plugins
  skill-creator/       # Fork of Anthropic's skill-creator
  skill-reviewer/      # Skill review tool
  skills-search/       # Skill discovery tool

daymade-claude-code/   # Claude Code-specific skills
  claude-code-history-files-finder/
  claude-md-progressive-disclosurer/
  continue-claude-work/
  ...

.claude-plugin/        # Claude plugin registry
  marketplace.json

scripts/               # Repository management scripts
.githooks/             # Pre-commit hooks (gitleaks integration)
.pre-commit-config.yaml
.gitleaks.toml
pyrightconfig.json

Security Infrastructure

  • .gitleaks.toml — secret scanning configuration
  • .pre-commit-config.yaml — pre-commit hook setup
  • scripts/security_scan.py (in skill-creator) — automated security gate
  • .githooks/ — pre-commit hooks

Required Runtime

  • Python 3.x (for automation scripts)
  • gitleaks (for security scanning in skill-creator)

Target AI Tools

Claude Code only.

03

Components

daymade/claude-code-skills — Components

Totals

Type Count
Skills 52
Plugin directories 45
Commands 0 (slash commands)
Subagents 0

Skill Categories (52 skills, representative)

Meta / Skill Development

  • skill-creator — fork of Anthropic's official skill-creator with 8-channel research + security gate
  • skill-reviewer — three-mode skill review (self, external, auto-PR)
  • skills-search — discover existing skills before building

Claude Code Specific (daymade-claude-code plugin, 7 skills)

  • claude-code-history-files-finder — find Claude context files in history
  • claude-md-progressive-disclosurer — implement progressive disclosure in CLAUDE.md
  • claude-export-txt-better — improve Claude conversation text exports
  • continue-claude-work — resume interrupted Claude sessions
  • statusline-generator — status line display generation
  • claude-skills-troubleshooting — debug skill activation issues
  • marketplace-dev — marketplace plugin development tools

QA / Testing

  • qa-expert — Google Testing Standards + OWASP with autonomous execution master prompt

Domain Skills (sampled from 45 plugin dirs)

  • capture-screen — screen capture skill
  • cli-demo-generator — CLI demo creation
  • cloudflare-troubleshooting — Cloudflare issue diagnosis
  • competitors-analysis — competitive analysis
  • deep-research — research methodology
  • excel-automation — Excel automation
  • fact-checker — fact verification
  • feishu-doc-scraper — Feishu (Lark) document scraping
  • financial-data-collector — financial data gathering
  • i18n-expert — internationalization expertise
  • iOS-APP-developer — iOS development
  • ima-copilot — IMA platform integration
  • llm-icon-finder — LLM icon discovery
  • macos-cleaner — macOS cleanup
  • product-analysis — product analysis
  • prompt-optimizer — prompt optimization
  • promptfoo-evaluation — Promptfoo eval framework
  • repomix-safe-mixer — Repomix with safety controls
  • scrapling-skill — web scraping
  • slides-creator — presentation creation
  • teams-channel-post-writer — Microsoft Teams integration
  • terraform-skill — Terraform infrastructure
  • tunnel-doctor — network tunnel debugging
  • twitter-reader — Twitter/X reading
  • ui-designer — UI design assistance
  • video-comparer — video comparison
  • youtube-downloader — YouTube content download
  • douban-skill — Douban (Chinese social platform) integration
  • gangtise-copilot — (Chinese-specific platform)
  • github-contributor / github-ops — GitHub workflow skills

Skill Structure (per skill)

Each skill follows the structure:

<skill-name>/
  SKILL.md           # Primary skill file
  scripts/           # Python automation
  references/        # Detailed methodology docs
  assets/            # Templates

The skill-creator skill includes:

  • quick_validate.py — fast structural validation
  • security_scan.py — gitleaks-based secret scanning
  • eval-viewer/generate_review.py — evaluation result visualization
  • references/evaluation_checklist.md — manual evaluation guide
  • references/skill-development-methodology.md — full methodology
05

Prompts

daymade/claude-code-skills — Prompt Excerpts

Excerpt 1: skill-creator SKILL.md (Anthropic fork)

Technique: AskUserQuestion pattern + progressive iteration loop (meta-skill authoring)

---
name: skill-creator
description: Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
---

# Skill Creator

At a high level, the process of creating a skill goes like this:
- Decide what you want the skill to do and roughly how it should do it
- Write a draft of the skill
- Create a few test prompts and run claude-with-access-to-the-skill on them
- Help the user evaluate the results both qualitatively and quantitatively

Your job when using this skill is to figure out where the user is in this process and then jump in and help them progress through these stages.

### Using AskUserQuestion (Critical — Read This)

**Use the AskUserQuestion tool aggressively at every decision point.**

Every AskUserQuestion MUST follow this structure:
1. **Re-ground**: State the skill name, current phase, and what just happened (1-2 sentences)
2. **Simplify**: Explain the decision in plain language. No function names or internal jargon.
3. **Recommend**: Lead with your recommendation and a one-line reason why.
4. **Options**: Provide 2-4 concrete, lettered choices.

Rules:
- One decision per question — never batch unrelated choices
- Accept the user's choice — nudge on tradeoffs but never refuse to proceed
- Skip the question if there's an obvious answer (just state what you'll do)

Analysis: AskUserQuestion structured invocation pattern is the most explicit UX design in any skill in this batch. Re-grounding instruction ("State the skill name, current phase") is a specific anti-context-loss technique.


Excerpt 2: qa-expert SKILL.md

Technique: One-command initialization + autonomous master prompt pattern

## Quick Start

**One-command initialization**:
```bash
python scripts/init_qa_project.py <project-name> [output-directory]

What gets created:

  • Directory structure (tests/docs/, tests/e2e/, tests/fixtures/)
  • Tracking CSVs (TEST-EXECUTION-TRACKING.csv, BUG-TRACKING-TEMPLATE.csv)
  • Documentation templates (BASELINE-METRICS.md, WEEKLY-PROGRESS-REPORT.md)
  • Master QA Prompt for autonomous execution

For autonomous execution (recommended): See references/master_qa_prompt.md — single copy-paste command for 100x speedup.


**Analysis**: "Master prompt" as a separate reference file (loaded on demand) implements progressive disclosure for complex workflows. The "100x speedup" claim for autonomous execution is a rare concrete performance benchmark in skill documentation.
09

Uniqueness

daymade/claude-code-skills — Uniqueness & Positioning

Differs From Seeds

Closest to spec-kit (quality-over-quantity, methodological depth) but focuses on skill craftsmanship rather than spec-driven development. Unlike superpowers (14 skills, behavioral Iron Laws), daymade is explicitly a practitioner's toolkit with quantitative evaluation. No seed has a skill whose purpose is to improve other skills (skill-creator fork, skill-reviewer). No seed targets the meta-skill creation process as a primary feature. The AskUserQuestion structured invocation protocol with explicit re-grounding, plain-language simplification, and recommendation pattern is a unique UX design not documented in any seed. The Chinese platform integrations (Feishu/Lark, Douban, Gangtise) indicate a Chinese developer community focus not present in any seed.

Distinctive Position

The only repo in this batch focused primarily on improving the skill-authoring process. The daymade skill-creator fork's quality benchmarks (42 vs 65 out of 80 across 8 dimensions) represent the most explicit quality improvement claim in the batch. The security scan gate (gitleaks) in skill creation is a unique safety mechanism.

Observable Failure Modes

  1. Eval infrastructure dependency: skill-creator depends on external eval infrastructure (running prompts, generating review) that requires Claude API access beyond normal Claude Code usage.
  2. skill-reviewer Mode 3 (auto-PR): requires GitHub authentication and repo permissions — likely to fail silently if not configured.
  3. Chinese platform skills: Feishu, Douban, Gangtise skills are only useful for users of those platforms — reduces general usability of the catalog.

Explicit Antipatterns (from skill-creator)

  • Open-ended questions when structured choices exist ("use AskUserQuestion aggressively")
  • Batching unrelated decisions in a single question
  • Refusing to proceed after user makes a choice
  • Building before researching existing skills (hence skills-search as prerequisite)
04

Workflow

daymade/claude-code-skills — Workflow

Skill Creation Workflow (skill-creator)

  1. User requests new skill: "Create a skill for X"
  2. 8-channel research protocol: search existing skills first (adopt/extend/build decision)
  3. 9 AskUserQuestion checkpoints: structured choices at every decision point (skill name, approach, scope, format, etc.)
  4. Draft skill → automated eval → LLM quality review → iterate
  5. Security scan gate: gitleaks check before packaging
  6. Description optimizer run: improves triggering accuracy

Skill Review Workflow (skill-reviewer, 3 modes)

Mode 1: Self-Review

  1. Auto-install skill-creator dependency
  2. Run quick_validate.py — structural check
  3. Run security_scan.py — secret detection
  4. Manual evaluation against checklist

Mode 2: External Review

  1. Evaluate someone else's skill repository
  2. Check against official best practices
  3. Provide improvement recommendations

Mode 3: Auto-PR

  1. Fork the target skill repository
  2. Improve the skill
  3. Submit pull request automatically

QA Expert Autonomous Mode

python scripts/init_qa_project.py <project-name>
# Creates: directory structure, tracking CSVs, templates, master QA prompt

Master QA Prompt: single copy-paste command for "100x speedup" autonomous execution.

Phase-to-Artifact Map

Phase Artifact
skill-creator research 8-channel research output
skill-creator draft SKILL.md draft
skill-creator eval Evaluation results + metrics
skill-creator security gate Security scan report
qa-expert init Test directory structure + tracking CSVs
06

Memory Context

daymade/claude-code-skills — Memory & Context

No persistent memory system. File-based context.

Skill evaluation runs create output files (eval results) as ephemeral session artifacts. No cross-session handoff protocol.

The continue-claude-work skill in daymade-claude-code plugin provides session resumption capabilities.

07

Orchestration

daymade/claude-code-skills — Orchestration

No multi-agent orchestration. Skills activate individually. No orchestrator patterns.

skill-reviewer Auto-PR Mode

The closest to orchestration: skill-reviewer Mode 3 forks a repo, improves a skill, and submits a PR — a three-step automated sequence.

Execution Mode

Interactive-loop. Skills activate contextually.

Cross-Tool Portability

Single-tool (Claude Code only).

08

Ui Cli Surface

daymade/claude-code-skills — UI / CLI Surface

No dedicated CLI or dashboard. Claude Code plugin marketplace only.

Python scripts within skills provide CLI tools (invoked by the skill, not by users directly):

  • init_qa_project.py — QA project initialization
  • quick_validate.py — skill structural validation
  • security_scan.py — gitleaks secret scanning
  • eval-viewer/generate_review.py — evaluation result display

Pre-commit hooks (.githooks/, .pre-commit-config.yaml) enforce quality on the repo itself.

Related frameworks

same archetype · same primary tool · same memory type

Superpowers ★ 207k

Enforces spec-first, TDD, and subagent-reviewed development as mandatory automatic workflows rather than optional practices.

Anthropic Skills (Official) ★ 141k

Official Anthropic reference gallery for the Agent Skills standard, spanning creative, technical, and enterprise capability…

OpenAI Skills (Official) ★ 20k

Tiered library of integration and workflow skills for Codex, with a self-bootstrapping skill-installer that lets agents discover…

Superpowers Lab ★ 344

Experimental skill incubator extending Superpowers with interactive CLI automation, on-demand MCP discovery, semantic…

openspec-skills (chyiiiiiiiiiiii) ★ 6

Ports the OpenSpec proposal/apply/archive SDD lifecycle into three plain SKILL.md files requiring zero runtime dependencies.

Anthropic Skills — IP Guard Delta Report ★ 141k

No ip-guard skill found in the anthropics/skills repo; this entry is a delta report tracking its absence.