Agent skill
skill-factory
Create new skills with proper structure and YAML frontmatter. Use when building new slash command skills, ensuring consistent formatting, directory structure, and validation. Guides through the complete skill creation workflow.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/skill-factory-euda1mon1a-autonomous-assignmen
SKILL.md
Skill Factory
Purpose: Guide users through creating new Claude Code skills with proper structure Created: 2025-12-27 Trigger:
/skill-factorycommand
When to Use
- Creating a new slash command skill
- Need to ensure proper YAML frontmatter format
- Want consistent skill structure across the project
- Building skills for agents or workflows
- Validating existing skill files
Required Actions
When this skill is invoked, Claude MUST:
-
Gather skill requirements from the user:
- Skill name (kebab-case, e.g.,
my-new-skill) - One-line description (for slash command discovery)
- Purpose and use cases
- Whether it needs Reference/ or Workflows/ subdirectories
- Skill name (kebab-case, e.g.,
-
Validate the skill name:
- Must be kebab-case
- Must not conflict with existing skills
- Must be descriptive and discoverable
-
Create the directory structure:
.claude/skills/<skill-name>/ ├── SKILL.md # Required: Main skill file ├── Reference/ # Optional: Reference documentation └── Workflows/ # Optional: Workflow definitions -
Generate SKILL.md using the template below
-
Validate the created skill:
- YAML frontmatter is valid
- Required sections are present
- Examples are included
Skill Template
Use this template for all new skills:
---
name: <skill-name>
description: <one-line description for slash command discovery>
---
# <Skill Title>
> **Purpose:** <what this skill does>
> **Created:** <date>
> **Trigger:** `/<skill-name>` command
---
## When to Use
<bullet list of scenarios when this skill should be used>
---
## Required Actions
When this skill is invoked, Claude MUST:
1. <action 1>
2. <action 2>
3. <action 3>
---
## Examples
<usage examples showing how the skill works>
---
## Escalation Rules
**Escalate to human when:**
<list of situations requiring human intervention>
**Can handle automatically:**
<list of situations the skill can handle independently>
---
## Related
- <related skills>
- <related documentation>
Validation Checklist
Before completing skill creation, verify:
-
YAML Frontmatter Valid
name:matches directory name (kebab-case)description:is one line, under 200 characters- No trailing spaces or invalid YAML syntax
-
Required Sections Present
- Title with Purpose/Created/Trigger metadata
- "When to Use" section with bullet points
- "Required Actions" section with numbered steps
- "Examples" section with concrete usage
-
Quality Standards
- Description is discoverable (keywords users would search)
- Actions are specific and actionable
- Examples demonstrate real use cases
- Escalation rules define boundaries
-
No Conflicts
- Skill name doesn't duplicate existing skill
- Functionality doesn't overlap significantly with existing skills
-
Directory Structure Correct
.claude/skills/<skill-name>/SKILL.mdexists- Optional subdirectories created if needed
Examples
Example 1: Create a Simple Skill
User: Create a skill for generating changelogs
Claude:
- Gathers requirements: name=
changelog-generator, purpose=generate changelogs from git history - Creates directory:
.claude/skills/changelog-generator/ - Generates SKILL.md with proper frontmatter
- Validates all checklist items pass
Example 2: Create a Skill with Reference Docs
User: Create a skill for ACGME compliance with reference documentation
Claude:
- Gathers requirements including reference materials needed
- Creates structure:
.claude/skills/acgme-compliance/ ├── SKILL.md └── Reference/ ├── hour-limits.md └── supervision-ratios.md - Populates reference files as needed
Example 3: Check Existing Skill
User: Validate the test-writer skill
Claude:
- Reads
.claude/skills/test-writer/SKILL.md - Validates YAML frontmatter
- Checks all required sections present
- Reports any issues found
Escalation Rules
Escalate to human when:
- Skill affects security (auth, credentials, secrets)
- Skill duplicates existing functionality significantly
- Skill requires new agent permissions
- Unclear whether skill or agent is appropriate
- Skill would modify critical system files
Can handle automatically:
- Creating standard skill structure
- Generating SKILL.md from template
- Validating existing skills
- Creating Reference/ and Workflows/ subdirectories
- Checking for naming conflicts
Integration with TOOLSMITH Agent
This skill implements part of the TOOLSMITH agent's "Create New Skill" workflow:
- TOOLSMITH receives skill creation request
- Invokes
/skill-factoryto generate structure - Validates output meets quality standards
- Reports completion to ORCHESTRATOR
For agent creation, use the /agent-factory skill instead.
Related
.claude/Agents/TOOLSMITH.md- Agent specification for tool creation.claude/skills/- Directory containing all project skillsdocs/development/AGENT_SKILLS.md- Agent skills referenceCLAUDE.md- Project guidelines and standards
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?