agentleFS
mikeparcewskimikeparcewski/wicked-garden/.claude/skills/scaffolding/SKILL.md

scaffold

Skill9 starsChanged 7 months agoMITView on GitHub

Component scaffolding toolkit for creating skills, context:fork worker skills, and hooks within the unified wicked-garden plugin. Generates complete, valid structures that pass validation out-of-the-box. Use when creating new domain components, setting up boilerplate, or ensuring proper structure from the start.

  • Deletes or force-pushes

50 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use ascii-diagram, docs-ci and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • SethGammonSkill

    ascii-diagram

    Generate perfectly aligned ASCII diagrams — architecture, flow, sequence, box-and-arrow. Uses a programmatic character-grid approach so alignment is guaranteed by math, not token prediction. Includes post-render verification.

    Also in 4 projects that use this

  • jh941213Skill

    docs-ci

    문서 CI(링크·OpenAPI lint·breaking 게이트·mermaid 검증)나 문서 드리프트 감지(docs.yaml 신선도 검사)를 설치하거나 돌릴 때 사용. 문서 내용 작성은 docs-architecture·docs-interfaces·docs-manuals.

    Also in 3 projects that use this

  • jh941213Skill

    docs-interfaces

    API·이벤트 인터페이스 문서(OpenAPI·swagger·AsyncAPI 명세, 엔드포인트 레퍼런스, API 구성도·시퀀스, API CHANGELOG)를 만들거나 코드와 맞출 때 사용. 시스템 구성도는 docs-architecture, API 설계 원칙은 api-design-principles.

    Also in 3 projects that use this

  • SethGammonSkill

    research-fleet

    Parallel research using Fleet wave mechanics. Spawns multiple scout agents, each investigating a different angle of the same question. Findings are compressed between waves. Produces a unified research brief from multiple independent perspectives.

    Also in 3 projects that use this

  • RachaelQuiselSkill

    client-docs

    Client documentation. Use when a build needs documentation updates or missing project documents need to be identified.

    Also in 3 projects that use this

  • vicoa-aiSkill

    live-preview

    Start a local app in the current user project, expose it through a tunnel, and return preview details in structured Markdown. Use when the user asks for a preview URL, public URL, shareable localhost, mobile/web preview, or tunnel setup.

    Also in 4 projects that use this

  • +4 more from jh941213

The file

We split this skill into the rules it gives an agent

10 rules. 0 are common, 4 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 43 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All scaffolded components pass validation immediately.

    line 26
  2. Distinctive:

    Use kebab-case for all names, max 64 characters

    line 245
  3. Very specific:

    No reserved prefixes (claude-code-, anthropic-, official-)

    line 246
  4. Very specific:

    Put reusable logic in scripts/

    line 251
  5. Very specific:

    Use skills/{domain}/ for the domain router + its actions (former commands)

    line 252
  6. Very specific:

    Use skills/{domain}-{role}/ for context:fork worker skills (former agents)

    line 253
  7. Very specific:

    Use hooks/ for deterministic event automation

    line 254
  8. Distinctive:

    Write clear, concise descriptions and include usage examples

    line 258
  9. Distinctive:

    Document dependencies and configuration options

    line 259
  10. Very specific:

    All scripts should use ${CLAUDE_PLUGIN_ROOT} for paths

    line 263

“Common” is measured: kept by at least 6 in 10 of the 43 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Scaffold Tool

Quick-start generator for wicked-garden components. The plugin is skills-only: the former commands/ and agents/ trees were absorbed into skills/, so this tool scaffolds skills, workers, and hooks — never a commands/ or agents/ file.

Purpose

Generates production-ready component structures within the unified plugin:

  • Skills — a sub-skill skills/{domain}/{name}/SKILL.md (Tier-2 navigation + refs/)
  • Worker skills — a standalone context: fork worker skills/{domain}-{role}/SKILL.md (the former "agent"), dispatched via Skill(skill="wicked-garden-{domain}-{role}")
  • Hooks — hooks.json entries with Python scripts

Former commands are now actions of the consolidated per-domain router skill (skills/{domain}/SKILL.md); the command subcommand is retired and only prints guidance for adding an action.

All scaffolded components pass validation immediately.

Usage

Interactive Mode

python .claude/skills/scaffolding/scripts/scaffold.py

# Interactive prompts:
# > Component type? (1 skill / 2 worker / 3 action-guidance / 4 hook)
# > Domain? (crew, engineering, platform, ...)
# > Name? (kebab-case)
# > Description?

# Result (skill):
# Skill created: skills/crew/my-skill/
# Skill id: my-skill (routed by the crew domain skill)

Command Line Mode

# Skill (a sub-skill under a domain)
python .claude/skills/scaffolding/scripts/scaffold.py skill \
  --name my-skill \
  --domain crew \
  --description "What this skill does" \
  --use-when "trigger conditions"

# Worker skill (context:fork — the former agent). `agent` is a back-compat alias.
python .claude/skills/scaffolding/scripts/scaffold.py worker \
  --name my-worker \
  --domain platform \
  --description "What this worker does" \
  --tools "Read,Write,Bash"

# Action (former command) — RETIRED. Prints guidance to add the action to the
# domain router skill; writes no file.
python .claude/skills/scaffolding/scripts/scaffold.py command \
  --name my-action \
  --domain engineering \
  --description "What this action does"

# Hook
python .claude/skills/scaffolding/scripts/scaffold.py hook \
  --event PreToolUse \
  --script validate-tool-use \
  --description "Validates tool usage"

Generated Paths

Components are placed in the skills-only layout at the repo root:

wicked-garden/                     # Plugin root
├── skills/{domain}/{name}/        # Sub-skills
│   └── SKILL.md
├── skills/{domain}-{role}/        # Worker skills (context:fork — former agents)
│   └── SKILL.md
├── hooks/scripts/{name}.py        # Hook scripts
└── hooks/hooks.json               # Hook bindings (updated)

There is no commands/ or agents/ tree.

Naming Conventions

  • Sub-skill id: the leaf directory name, routed by its domain skill
  • Worker skill name: wicked-garden-{domain}-{role} (dash-separated), declared in frontmatter name: and dispatched via Skill(skill="…")
  • Domain router skill name: wicked-garden-{domain}

Action stub (former command-with-agent)

A former "command that dispatched to an agent" is now an action of the domain router skill that dispatches to a worker skill. Copy the stub in templates/plugin/command-with-agent.md into the router body and wire it with:

Skill(skill="wicked-garden-{domain}-{role}", args="… | target: <target>")

Skill Template

---
name: {skill-name}
description: |
  What this skill does (capabilities).
  Use when [trigger conditions and keywords].
---

# {Skill Title}

Brief introduction to the skill's purpose.

## Purpose
What problem this skill solves.

## Usage
Examples.

## References
- Related docs / source files

Worker Skill Template (context:fork)

---
name: wicked-garden-{domain}-{role}
description: What this worker specializes in
context: fork
allowed-tools: ["Read", "Write", "Bash"]
model: sonnet
color: blue
---

# {Worker Title}

You are the {Worker Title} worker, specialized in {domain}. You run in an
isolated context:fork subagent, dispatched via Skill(skill="…").

## Expertise
- Capability 1 / 2 / 3

## Working Style
1. Step 1 / 2 / 3

## Quality Standards
- Standard 1 / 2

## Constraints
- Constraint 1 / 2

context: fork worker skills are exempt from the ≤200-line Tier-2 cap — they load into an isolated subagent context, not the parent.

Hook Template

hooks.json

{
  "hooks": [
    {
      "event": "{event}",
      "script": "scripts/{script-name}.py",
      "description": "{description}",
      "enabled": true
    }
  ]
}

Hook Script

#!/usr/bin/env python3
"""
{Event} hook for {plugin-name}.

Exit codes:
  0 - Success, continue
  2 - Blocking error (message sent to Claude)
  Other - Non-blocking error (logged)
"""

import sys
import json

def main():
    try:
        data = json.loads(sys.stdin.read())
    except json.JSONDecodeError:
        print("Error: Invalid JSON input", file=sys.stderr)
        sys.exit(1)

    # Hook logic here. Access: data['tool'], data['arguments'], data['context']
    if data.get('tool') == 'Bash':
        command = data.get('arguments', {}).get('command', '')
        if 'rm -rf /' in command:
            print("Blocked dangerous command: rm -rf /")
            sys.exit(2)  # Block execution

    sys.exit(0)

if __name__ == "__main__":
    main()

Valid Domains

Domains are discovered dynamically from .claude-plugin/components.json ("domains"), falling back to the top-level directories under skills/ (excluding the {domain}-{role} worker dirs). New domains are valid as soon as they appear in the manifest / skills tree.

Current domains: agentic, crew, data, engineering, jam, mem, persona, platform, product, qe, search.

Template Variables

VariableDescriptionExample
{{name}}Component name (kebab-case)risk-assessor
{{title}}Component name (Title Case)Risk Assessor
{{skill_name}}Worker skill dash namewicked-garden-qe-risk-assessor
{{description}}Brief descriptionAssess project risks
{{domain}}Area of expertise / domainqe
{{tools}}Allowed-tools list (rendered)"Read", "Write"

Best Practices

Naming

  • Use kebab-case for all names, max 64 characters
  • No reserved prefixes (claude-code-, anthropic-, official-)
  • Worker skills carry the dash-qualified wicked-garden-{domain}-{role} name

Structure

  • Put reusable logic in scripts/
  • Use skills/{domain}/ for the domain router + its actions (former commands)
  • Use skills/{domain}-{role}/ for context:fork worker skills (former agents)
  • Use hooks/ for deterministic event automation

Documentation

  • Write clear, concise descriptions and include usage examples
  • Document dependencies and configuration options

Security

  • All scripts should use ${CLAUDE_PLUGIN_ROOT} for paths
  • Quote shell variables: "$VAR"
  • Validate input before processing; no hardcoded secrets

Examples

Sub-skill only

python .claude/skills/scaffolding/scripts/scaffold.py skill \
  --name validation-patterns \
  --domain engineering \
  --description "Input validation patterns" \
  --use-when "reviewing input handling"

# Result:
# skills/engineering/validation-patterns/SKILL.md

Worker skill (context:fork)

python .claude/skills/scaffolding/scripts/scaffold.py worker \
  --name risk-assessor \
  --domain qe \
  --description "Assesses release risk" \
  --tools "Read,Bash"

# Result:
# skills/qe-risk-assessor/SKILL.md   (name: wicked-garden-qe-risk-assessor, context: fork)
# Dispatch via: Skill(skill="wicked-garden-qe-risk-assessor")

Hook with Script

python .claude/skills/scaffolding/scripts/scaffold.py hook \
  --event PreToolUse \
  --script validate-tool-use \
  --description "Validate tool use before execution"

# Result:
# hooks/hooks.json (updated)
# hooks/scripts/validate-tool-use.py

References

  • Naming conventions: .claude/CLAUDE.md (Naming Conventions section)
  • Component manifest: .claude-plugin/components.json (domains + skills)
  • Cross-tool context: Plugins that read project descriptor files should load AGENTS.md before CLAUDE.md (general → specific). AGENTS.md is read-only.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.