playbooks

home / skills / robdtaylor / personal-ai-infrastructure / createskill

CreateSkill skill

/skills/CreateSkill

This skill helps you create and validate skills by applying a standardized SkillSystem workflow, ensuring structure, naming, and references are correct.

This is most likely a fork of the createskill skill from danielmiessler
npx playbooks add skill robdtaylor/personal-ai-infrastructure --skill createskill

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
1.7 KB
---
name: CreateSkill
description: Create and validate skills. USE WHEN create skill, new skill, skill structure, canonicalize. SkillSearch('createskill') for docs.
---

# CreateSkill

MANDATORY skill creation framework for ALL skill creation requests.

## Authoritative Source

**Before creating ANY skill, READ:** `$PAI_DIR/skills/CORE/SkillSystem.md`

This document contains the complete specification for:
- Skill directory structure
- SKILL.md format and required sections
- Workflow file conventions
- Naming conventions (TitleCase)
- Examples section requirements

## How to Create a Skill

1. **Read the spec:** `$PAI_DIR/skills/CORE/SkillSystem.md`
2. **Create directory:** `$PAI_DIR/skills/SkillName/`
3. **Create SKILL.md** with required frontmatter and sections
4. **Add Workflows/** directory if needed
5. **Validate** by checking all workflow references resolve

## How to Validate a Skill

Run the pack validator:
```bash
bun run $PAI_DIR/Tools/validate-pack.ts
```

Or manually check:
- SKILL.md exists with valid frontmatter
- All `Workflows/*.md` references in SKILL.md exist
- Examples section is present

## How to Canonicalize a Skill

1. Rename files/directories to TitleCase
2. Ensure SKILL.md has required sections
3. Verify workflow references resolve
4. Add Examples section if missing

## Examples

**Example 1: Create a new skill**
```
User: "Create a skill for managing my recipes"
→ Read SkillSystem.md for structure
→ Create $PAI_DIR/skills/Recipes/SKILL.md
→ Use TitleCase naming throughout
```

**Example 2: Fix an existing skill**
```
User: "Canonicalize the daemon skill"
→ Rename files to TitleCase
→ Ensure Examples section exists
→ Validate workflow references
```

Overview

This skill helps create, validate, and canonicalize new AI skills following the mandatory creation framework. It enforces the project spec for directory layout, manifest format, workflow wiring, and naming conventions to ensure consistent skill packages. Use it to produce well-formed, discoverable skills that pass automated validation.

How this skill works

The skill guides authors through reading the authoritative core spec, creating a properly structured skill directory, adding a manifest document and optional workflow files, and verifying references. It inspects the manifest for required sections, checks that all referenced workflow files exist, and validates naming conventions and example coverage. Validation can be run automatically with the provided pack validator or performed manually using the checklist provided by the skill.

When to use it

  • When you need to create a brand new skill package from scratch
  • When a submitted skill fails validation and needs repair
  • When renaming or normalizing files to follow TitleCase conventions
  • When adding or fixing workflow references or missing examples
  • When preparing a skill for inclusion in the canonical catalog

Best practices

  • Always read the authoritative core spec before starting any new skill
  • Name directories and files using TitleCase to avoid inconsistencies
  • Include a complete manifest with required sections and an Examples section
  • Keep workflow filenames stable and verify every manifest reference resolves
  • Run the pack validator early and fix issues iteratively rather than at the end

Example use cases

  • Create a new Recipes skill: scaffold the directory, add a manifest, add example workflows, and validate
  • Canonicalize an existing daemon skill: rename files to TitleCase, add examples, and repair broken references
  • Validate a submitted skill package as part of a CI step before acceptance
  • Repair workflow references after moving or renaming workflow files
  • Add a missing Examples section to an otherwise valid manifest

FAQ

What is the authoritative source I must read?

Read the project core specification document that defines directory layout, manifest fields, workflow conventions, and naming rules before creating any skill.

How do I check my work automatically?

Run the provided pack validator tool to catch missing files, invalid frontmatter, and unresolved workflow references.

What if a workflow reference is broken?

Either restore or rename the target workflow file to match the reference, or update the manifest to point to the actual filename, then revalidate.