Skill Creator

For Claude: How to Use This Skill

Skill instructions are preloaded in context when this skill is active. Do not perform filesystem searches or tool-driven exploration to rediscover them; use the guidance below directly.

Use this skill to create new skills or modernize existing ones in this repository.

When To Use

Use this skill when the user asks to:

• create a new skill
• update skill frontmatter or layout
• make a skill compatible with Agent Skills-compliant tooling

Canonical Standard

Treat the Agent Skills specification as the source of truth:

• https://agentskills.io/specification

Core spec rules:

• SKILL.md must include YAML frontmatter
• required top-level fields: name, description
• optional top-level fields: license, compatibility, metadata, allowed-tools
• name must:
  • be 1-64 chars
  • use lowercase letters, digits, and hyphens only
  • not start/end with a hyphen
  • not contain consecutive hyphens
  • exactly match the parent directory name
• if metadata is used, it must be a map of string keys to string values

Seren Repo Conventions

On top of the spec, this repo uses the following conventions:

• frontmatter uses only name and description (everything else is derived)
• optional spec fields (license, compatibility, allowed-tools, metadata) are available when appropriate
• use the first # H1 in the body as the display name
• keep runtime files in scripts/
• keep requirements.txt / package.json, config.example.json, and .env.example at the skill root
• keep local config.json untracked

Workflow

1. Confirm Scope

Collect:

• org name (for example seren, coinbase)
• skill directory name in kebab-case (for example grid-trader)
• skill purpose and activation context
• whether the skill needs executable code (put in scripts/) or is documentation-only

2. Scaffold Directory

Create:

• {org}/{skill-name}/SKILL.md
• optional runtime files under {org}/{skill-name}/scripts/
• optional root files: requirements.txt, package.json, config.example.json, .env.example

3. Write Frontmatter Correctly

Use this baseline template:

---
name: skill-name
description: Clear statement of what the skill does and when to use it
# license: Apache-2.0
# compatibility: "Requires git and jq"
# allowed-tools: Read Bash(git:*)
---

Validation requirements:

• name equals directory name exactly
• description is concrete and trigger-oriented
• optional fields (if present) follow spec constraints
• keep frontmatter fields minimal and relevant

4. Write A Focused SKILL.md Body

Include:

• when to use the skill
• step-by-step workflow
• minimal command examples
• constraints and failure handling

Avoid:

• long generic prose
• duplicate documentation across many files
• references to paths that do not exist

5. Validate Before PR

Checklist:

• frontmatter present and valid
• name matches directory
• commands in docs point to real files
• runtime files (if any) live under scripts/
• no secrets committed

Example Skeleton

---
name: api-tester
description: Test REST endpoints and validate response contracts.
---

# API Tester

## When to Use

Use when a user asks to validate API behavior or contracts.

## Workflow

1. Collect endpoint and auth details.
2. Run a defined test matrix (happy path + failures).
3. Report findings with reproducible commands.