Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Workflow Creator · nicepkg | DecimalAI

nicepkg/workflow-creator

SkillScore83Maintained

Create complete Claude Code workflow directories with curated skills. Use when user wants to (1) create a new workflow for specific use case (media creator, developer, marketer, etc.), (2) set up a Claude Code project with pre-configured skills, (3) download and organize skills from GitHub repositories, or (4) generate README.md and AGENTS.md documentation for workflows. Triggers on phrases like "create workflow", "new workflow", "set up workflow", "build a xxx-workflow".

agent-toolingby nicepkg

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

AI Engineer pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/nicepkg-workflow-creator/SKILL.md

decimalai skills pull nicepkg-workflow-creator --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/nicepkg-workflow-creator/SKILL.md https://app.decimal.ai/s/nicepkg-workflow-creator/SKILL.md

Raw file: /s/nicepkg-workflow-creator/SKILL.md ↗ · machine-readable summary at /s/nicepkg-workflow-creator.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 5d ago

—+86% tokens0% turns83% of whole cases pass with skill (19/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-04	✗→✓	▲ Improved	182%	0%
case-18	✗→✓	▲ Improved	153%	0%
case-05	✗→✓	▲ Improved	175%	0%
case-06	✗→✓	▲ Improved	309%	0%
case-07	✗→✓	▲ Improved	154%	0%

Showing 5 of 23 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

nicepkg / ai-workflow

.claude/skills/workflow-creator/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1Bundled 3 files · scripts

Skill contents

scripts/

references/

SKILL.md

Workflow Creator

Create complete workflow directories with curated skills downloaded from GitHub.

Workflow Creation Process

Step 1: Create directory structure

Run scripts/create_workflow.py to initialize:

bash
python scripts/create_workflow.py <workflow-name> --path <output-dir>

Creates (with multi-AI tool support):

workflows/<workflow-name>-workflow/
├── README.md          # User documentation (English)
├── README_cn.md       # User documentation (Chinese)
├── AGENTS.md          # AI context (auto-loaded)
├── .claude/
│   ├── settings.json
│   └── skills/        # Skills go here (primary storage)
├── .codex/
│   └── skills -> ../.claude/skills
├── .cursor/
│   └── skills -> ../.claude/skills
├── .opencode/
│   └── skill -> ../.claude/skills
├── .agents/
│   └── skills -> ../.claude/skills
├── .kilocode/
│   └── skills -> ../.claude/skills
├── .roo/
│   └── skills -> ../.claude/skills
├── .goose/
│   └── skills -> ../.claude/skills
├── .gemini/
│   └── skills -> ../.claude/skills
├── .agent/
│   └── skills -> ../.claude/skills
├── .github/
│   └── skills -> ../.claude/skills
├── skills -> .claude/skills
├── .factory/
│   └── skills -> ../.claude/skills
└── .windsurf/
    └── skills -> ../.claude/skills

Symlinks enable all AI tools to use the same skills from .claude/skills/.

Step 2: Select and download skills

Use the Skill Sources section below to find relevant skills
Download each skill using scripts/download_skill.py:

bash
python scripts/download_skill.py <repo-url> <skill-path> --output workflows/<workflow-name>/.claude/skills/

Examples:

bash
# Official Anthropic skills
python scripts/download_skill.py https://github.com/anthropics/skills skills/docx --output ./workflows/media-workflow/.claude/skills/

# Community skills (root level)
python scripts/download_skill.py https://github.com/gked2121/claude-skills social-repurposer --output ./workflows/media-workflow/.claude/skills/

Step 3: Generate README.md (English)

CRITICAL: Follow the exact format from workflows/marketing-pro-workflow/README.md

Required structure:

markdown
<div align="center">

# 📋 Workflow Name

### **Your AI-Powered [Domain] Team**

[← Back to AI Workflow](../../README.md)

[简体中文](./README_cn.md) | English

</div>

---

## 🎯 Who Is This For?
- **Role 1** - Use case
- **Role 2** - Use case

---

## ⚡ Quick Install
[Install commands]

---

## 📦 Skills Included (N)

### 0️⃣ Stage Name
| Skill | What It Does |
|:------|:-------------|
| `skill-name` | Description |

### 1️⃣ Next Stage
...

---

## 🔄 Complete Pipeline (N Stages)
[ASCII tree diagram]

---

## 💡 Example Workflows
[Multiple scenario examples with numbered prompts]

---

## 🔗 Skill Combinations
[Table with Goal → Skill Chain]

---

## 📄 License
MIT © [nicepkg](https://github.com/nicepkg)

Use assets/templates/README.template.md as reference.

Step 4: Generate README_cn.md (Chinese)

Create Chinese version with:

Same structure as English README
Professional Chinese terminology
Link back to ../../README_cn.md

Step 5: Generate AGENTS.md

Write AI instructions covering:

Workflow overview with pipeline diagram
Available skills grouped by stage
Skill usage guidelines (when to use each)
Recommended sequences for common tasks
Output standards
Quality gates between phases

Important: AGENTS.md is auto-loaded by Claude Code. Keep it concise (<500 lines) and focused on actionable instructions.

Use assets/templates/AGENTS.template.md as reference.

Step 6: Update project README

After creating a workflow in the ai-workflow project, update:

README.md (English):

Add new workflow to the workflow table
Add skills list under <details> section
Update skill count (e.g., "150+ skills")

README_cn.md (Chinese):

Same updates in Chinese
Link to README_cn.md in workflow folder

Skill Sources

A. Skill Aggregators & Directories (for discovering skills)

| Name | Type | Scale | Best For | |:-----|:-----|------:|:---------| | Skillhub Awesome Skills | GitHub curated list | 1000+ skills | Top skills by category, well-organized | | Skillhub.club | Online directory | 1000+ | Web UI for browsing by category | | SkillsMP | Marketplace | 63,000+ | Mass search with filters (category/popularity/author) | | agent-skills.md | Online directory | Large | Install commands included (pnpm dlx add-skill ...) | | Claude Skills Hub | Online directory | Medium | Product-style browsing, good for inspiration | | MCP Market Skills | Online store | Medium | Product pages with About/FAQ, content creator friendly | | Skills Directory | Online directory | Medium | Copy SKILL.md directly, good structure | | Smithery Skills | Skill directory | Medium | Includes expected_output.json, tool-oriented | | Awesome Claude Skills (ComposioHQ) | GitHub awesome list | High stars | Ecosystem resources, comprehensive | | Awesome Claude Skills (travisvn) | GitHub awesome list | 5k+ stars | Second perspective, prevents single-source bias | | Awesome Agent Skills | GitHub multi-agent list | 1k+ stars | Covers Claude/Codex/Copilot/VSCode |

B. Production-Ready Skill Repositories (for bulk import)

| Repository | Focus | Best For | |:-----------|:------|:---------| | alirezarezvani/claude-skills | Multi-role skill packs | Content Creator suite (brand voice, SEO, platform frameworks) | | gked2121/claude-skills | Production workflows | Clear workflow examples (Podcast→Content→Social→SEO) | | sickn33/antigravity-awesome-skills | Skill registry | Organized by role (skills/content-creator) | | Microck/ordinary-claude-skills | 600+ skills collection | skill-navigator, SEO clustering, email optimization | | synapz-org/marketing-ops-hub | Marketing orchestration | Multi-skill coordination patterns | | m2ai-portfolio/claude-skills | Creator/Growth suite | IG caption, YouTube script, tweet thread, blog outline | | pluginagentmarketplace/custom-plugin-product-manager | Product Manager suite | user-research, roadmap, requirements, analytics | | lyndonkl/claude | Decision frameworks | prioritization, forecasting, stakeholder mapping | | jamesrochabrun/skills | Mixed professional | prd-generator, technical-launch-planner | | britt/claude-code-skills | Writing & specs | writing-product-specs, writing-user-stories | | troykelly/codex-skills | Workflow management | work-intake, milestone-management | | escarti/agentDevPrompts | Feature lifecycle | feature-planning, feature-implementing | | aj-geddes/useful-ai-prompts | Sprint & agile | agile-sprint-planning, requirements-gathering | | daffy0208/ai-dev-standards | Multi-role standards | customer-feedback-analyzer, go-to-market-planner |

C. Search Strategy

When the above sources lack needed skills:

GitHub Search: "claude" "skills" "SKILL.md" <topic>
agent-skills.md Search: Browse by tags at https://agent-skills.md/tags
SkillsMP Search: Filter by category at https://skillsmp.com
Validate: Check SKILL.md has valid YAML frontmatter with name and description

D. Download Commands

bash
# From GitHub repository
python scripts/download_skill.py https://github.com/<owner>/<repo> <skill-path> --output workflows/<workflow-name>/.claude/skills/

# Example: Download from alirezarezvani
python scripts/download_skill.py https://github.com/alirezarezvani/claude-skills skills/content-creator --output ./workflows/content-creator-workflow/.claude/skills/

# Example: Download from pluginagentmarketplace
python scripts/download_skill.py https://github.com/pluginagentmarketplace/custom-plugin-product-manager skills/user-research --output ./workflows/product-manager-workflow/.claude/skills/

Output Checklist

After workflow creation, verify:

] .claude/skills/ contains downloaded skill folders
] Each skill folder has SKILL.md
] All symlinks work (.codex/skills, .cursor/skills, .opencode/skill, .agents/skills, .kilocode/skills, .roo/skills, .goose/skills, .gemini/skills, .agent/skills, .github/skills, skills, .factory/skills, .windsurf/skills)
] README.md follows standard format with all sections
] README_cn.md created with Chinese translation
] AGENTS.md provides clear AI instructions (<500 lines)
] settings.json exists in .claude/
] Project README.md updated with new workflow
] Project README_cn.md updated with new workflow
] Skill counts updated in both README files

Want to see if nicepkg/workflow-creator would activate on your phrasing? Try it before you install — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/24/2026, 00:29 UTC · 23 cases

Measured on 20 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

+52%

83% vs 30%

Agent turns

1.0 vs 1.0

Tokens

+86%

3,332.3 vs 1,789.9

fail → pass · 13pass → pass · 6fail → fail · 3pass → fail · 1

Per-case results (23)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-17 When manually validating a newly discovered SKILL.md file from GitHub, what two required YAML frontmatter fields must be present?	pass→pass	5,004	1,323	-74%	1	1	0%	731	2,936	+302%	0	0	—
case-01 I need to build a brand new workflow package named 'devops-incident-workflow' in my AI workflows repository. Please set up the full directory structure, fetch relevant incident management and monitoring skills from GitHub into the primary storage folder, generate both English and Chinese README documentation following our standardized layout, draft the AGENTS context file with step-by-step pipeline guidance for AI assistants, and update the main project READMEs with the new workflow details.	fail→fail	24,122	19,724	-18%	1	1	0%	4,511	3,219	-29%	0	0	—
case-02 Could you help me assemble a complete product management workflow directory called 'product-management-workflow'? Please initialize the necessary folder setup and tool links, pull in PRD, user research, and roadmap skills from GitHub source repos into the skills folder, write the user documentation in both English and Simplified Chinese, create the AGENTS.md file outlining stage-by-stage instructions and quality standards, and register this new addition in the main repository's root documentation files.	fail→fail	43,057	7,473	-83%	1	1	0%	6,060	3,480	-43%	0	0	—
case-03 We are introducing a 'content-creator-workflow' to our repository. Please create the entire workflow scaffold, download curated writing and social media repurposing skills from GitHub into the proper skills directory, produce the English README along with its Chinese translation, prepare the AGENTS guide explaining how the AI should execute each stage of the pipeline, and make sure the global project README files are updated to reflect the newly added workflow and skill metrics.	fail→fail	29,032	5,129	-82%	1	1	0%	5,724	3,064	-46%	0	0	—
case-04 What exact command should be run to initialize a new workflow directory named 'data-engineering-workflow' under the 'workflows' folder?	fail→pass	7,016	3,135	-55%	1	1	0%	1,168	3,292	+182%	0	0	—
case-18 When completing a workflow addition, which root files in the main repository need skill count and workflow registry updates?	fail→pass	8,709	2,854	-67%	1	1	0%	1,282	3,247	+153%	0	0	—
case-05 When setting up multi-tool compatibility for AI workflow skills in this repo, where must downloaded skill directories be physically stored before creating symlinks?	fail→pass	7,640	2,678	-65%	1	1	0%	1,179	3,240	+175%	0	0	—
case-06 I am adding translated documentation for 'seo-pro-workflow'. What exact filename should be used for the Chinese version of the workflow README?	fail→pass	4,481	1,781	-60%	1	1	0%	740	3,026	+309%	0	0	—
case-07 Show the command line command to download the 'docx' skill from the official anthropic skills repository into 'media-workflow'.	fail→pass	8,047	1,892	-76%	1	1	0%	1,212	3,080	+154%	0	0	—
case-08 When authoring the AGENTS.md context guide for a complex 8-stage workflow, what maximum line count limit should be observed to avoid prompt bloat?	pass→pass	11,311	5,707	-50%	1	1	0%	1,695	3,695	+118%	0	0	—
case-09 How should the list of skills contained within a newly added workflow be rendered inside the main project README.md file?	pass→pass	12,705	6,425	-49%	1	1	0%	2,033	3,761	+85%	0	0	—
case-10 In the workflow directory structure, what relative path target should the symlink inside `.cursor/skills` point to?	fail→pass	11,387	2,100	-82%	1	1	0%	1,778	3,109	+75%	0	0	—
case-11 How should the top section of a workflow README.md be aligned and formatted?	pass→pass	11,984	4,000	-67%	1	1	0%	2,028	3,489	+72%	0	0	—
case-12 During the final verification step of workflow creation, what file must exist inside every downloaded skill subfolder to confirm it is valid?	fail→pass	13,805	1,552	-89%	1	1	0%	2,161	2,985	+38%	0	0	—
case-13 In the standard workflow scaffold, what is the exact name of the symlinked entry created inside `.opencode/`?	fail→pass	13,120	2,039	-84%	1	1	0%	2,126	3,086	+45%	0	0	—
case-14 Where can I search a high-volume marketplace containing tens of thousands of agent skills with category filters?	fail→pass	14,266	2,759	-81%	1	1	0%	2,082	3,225	+55%	0	0	—
case-15 What copyright owner attribution line should be placed in the License section of generated workflow README files?	fail→pass	8,834	1,723	-80%	1	1	0%	1,330	3,030	+128%	0	0	—
case-16 If skill aggregators do not yield the required skill, what exact search string structure should be used on GitHub to locate matching skills?	fail→pass	12,400	2,139	-83%	1	1	0%	1,859	3,071	+65%	0	0	—
case-19 Which specific AI assistant tool automatically loads AGENTS.md upon initialization in the project directory?	fail→pass	19,589	2,074	-89%	1	1	0%	3,326	3,098	-7%	0	0	—
case-20 Which production repository source is recommended for finding sprint planning and agile requirement gathering prompts?	fail→pass	13,382	2,792	-79%	1	1	0%	1,940	3,288	+69%	0	0	—
case-21 I have a bug in `scripts/download_skill.py` where requests timeout on large repositories. Please rewrite the python request handling code using `httpx` with retry backoff.	pass→fail	17,950	1,541	-91%	1	1	0%	3,393	2,962	-13%	0	0	—
case-22 Write a standalone prompt file for a new individual skill named 'pdf-text-extractor' that parses PDF documents using pdfplumber.	pass→pass	13,331	13,231	-1%	1	1	0%	2,237	5,173	+131%	0	0	—
case-23 Can you draft the git commit message and branch creation command for publishing changes to our GitHub remote repository?	pass→pass	9,514	7,714	-19%	1	1	0%	1,498	3,853	+157%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 20 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +52 percentage points is the difference between those two pass rates over the 20 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 30.4%
Pass rate with the skill: 82.6%
Cases attempted: 23
Cases counted toward lift: 20
Cases that regressed: 1
Skill version measured: v1
Run completed: 8/24/2026, 00:29 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-04

Prompt

What exact command should be run to initialize a new workflow directory named 'data-engineering-workflow' under the 'workflows' folder?

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run83

Live evalNeeds ≥20 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+124% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 5 days ago

Other measured skills in the registry, with their headline benchmark lift.