Back to list
arbgjr

skill-io-creator

by arbgjr

Sistema de desenvolvimento de software orientado por agentes de IA que automatiza e coordena todo o ciclo de vida do desenvolvimento.

1🍴 0📅 Jan 25, 2026

SKILL.md


name: skill-io-creator description: Create skills using the agentskills.io framework with init, validate, and package scripts. Use when users want to create distributable .skill packages, use the init_skill.py scaffolding, or follow the agentskills.io workflow with scripts, references, and assets directories. allowed-tools:

  • Read
  • Bash
  • Write
  • Edit

Skill IO Creator (agentskills.io Framework)

This skill provides the agentskills.io framework for creating distributable skill packages with initialization scripts, validation, and packaging tools.

Note: This is the agentskills.io framework, separate from the official Claude Code skills format. Use skill-creator for official Claude Code skills.

When to Use This vs skill-creator

Use CaseUse This (skill-io-creator)Use skill-creator
Distributable .skill packagesYesNo
Script-based scaffoldingYesNo
Official Claude Code formatNoYes
Frontmatter with hooks/contextNoYes

Framework Overview

The agentskills.io framework provides:

  1. init_skill.py - Scaffold new skill directories with templates
  2. quick_validate.py - Validate skill structure and content
  3. package_skill.py - Package skills into distributable .skill files

Skill Structure

skill-name/
├── SKILL.md                 # Required: frontmatter + instructions
├── scripts/                 # Executable code
│   └── helper.py
├── references/              # Documentation loaded into context
│   └── api_reference.md
└── assets/                  # Files used in output (templates, images)
    └── template.pptx

Creating a New Skill

Step 1: Initialize

python scripts/init_skill.py <skill-name> --path <output-directory>

Example:

python scripts/init_skill.py pdf-editor --path ~/.claude/skills

This creates:

  • SKILL.md with TODO placeholders
  • scripts/example.py - sample executable
  • references/api_reference.md - sample documentation
  • assets/example_asset.txt - sample asset placeholder

Step 2: Edit SKILL.md

Complete the frontmatter:

---
name: skill-name
description: What it does and WHEN to use it. Include trigger scenarios.
---

Choose a structure pattern:

PatternBest ForExample
Workflow-BasedSequential processesPDF: Analyze -> Map -> Fill -> Verify
Task-BasedTool collectionsPDF: Merge, Split, Extract
Reference/GuidelinesStandardsBrand: Colors, Typography
Capabilities-BasedIntegrated featuresPM: numbered capabilities

Step 3: Add Resources

scripts/ - Executable code

  • Python scripts, shell scripts
  • Must be tested before packaging
  • Can be executed without loading into context

references/ - Documentation

  • API docs, schemas, policies
  • Loaded into context when needed
  • Keep SKILL.md lean by moving details here

assets/ - Output files

  • Templates, images, fonts
  • NOT loaded into context
  • Copied/used in final output

Step 4: Validate

python scripts/quick_validate.py <path/to/skill-folder>

Checks:

  • YAML frontmatter format
  • Required fields (name, description)
  • Directory structure
  • Naming conventions

Step 5: Package

python scripts/package_skill.py <path/to/skill-folder> [output-dir]

Creates skill-name.skill file (zip with .skill extension).

Core Principles

1. Concise is Key

Context window is shared. Only add what Claude doesn't already know.

2. Progressive Disclosure

  1. Metadata (always loaded) - name + description
  2. SKILL.md body (on trigger) - instructions
  3. Resources (on demand) - scripts, references, assets

Keep SKILL.md under 500 lines.

3. Degrees of Freedom

LevelWhenExample
HighMultiple valid approachesText guidelines
MediumPreferred patternPseudocode with params
LowFragile operationsSpecific scripts

What NOT to Include

  • README.md
  • INSTALLATION_GUIDE.md
  • CHANGELOG.md
  • User documentation
  • Setup procedures

Design Patterns

Workflow Patterns

For sequential processes, provide overview:

1. Analyze the form (run analyze_form.py)
2. Create field mapping (edit fields.json)
3. Validate mapping (run validate_fields.py)
4. Fill the form (run fill_form.py)

For conditional workflows:

1. Determine type:
   **Creating?** → Follow Creation workflow
   **Editing?** → Follow Editing workflow

See references/workflows.md for more patterns.

Output Patterns

For strict output format:

ALWAYS use this exact template:

# [Title]
## Executive summary
## Key findings
## Recommendations

For flexible output:

Here is a sensible default, adapt as needed:
...

See references/output-patterns.md for more patterns.

Example: Creating a PDF Editor Skill

# 1. Initialize
python scripts/init_skill.py pdf-editor --path ~/.claude/skills

# 2. Edit SKILL.md
# - Update description with trigger scenarios
# - Add workflow steps
# - Reference scripts

# 3. Add scripts
# - scripts/rotate_pdf.py
# - scripts/extract_text.py

# 4. Add references
# - references/pdf_api.md

# 5. Validate
python scripts/quick_validate.py ~/.claude/skills/pdf-editor

# 6. Package
python scripts/package_skill.py ~/.claude/skills/pdf-editor ./dist

Script Usage Reference

init_skill.py

Usage: init_skill.py <skill-name> --path <path>

Skill name requirements:
  - Hyphen-case (e.g., 'data-analyzer')
  - Lowercase letters, digits, hyphens only
  - Max 40 characters

quick_validate.py

Usage: quick_validate.py <path/to/skill-folder>

Validates:
  - Frontmatter syntax
  - Required fields
  - Directory structure

package_skill.py

Usage: package_skill.py <path/to/skill-folder> [output-directory]

Creates .skill file (zip format) for distribution.

Score

Total Score

70/100

Based on repository quality metrics

SKILL.md

SKILL.mdファイルが含まれている

+20
LICENSE

ライセンスが設定されている

+10
説明文

100文字以上の説明がある

+10
人気

GitHub Stars 100以上

0/15
最近の活動

1ヶ月以内に更新

+10
フォーク

10回以上フォークされている

0/5
Issue管理

オープンIssueが50未満

0/5
言語

プログラミング言語が設定されている

+5
タグ

1つ以上のタグが設定されている

+5

Reviews

💬

Reviews coming soon