Skill Creator
Create, refine, score, and benchmark custom agent skills when a repeatable workflow deserves its own SKILL.md instead of one-off chat instructions.
Overview
Skill Creator is a journey-wide agent skill that creates, improves, and measures SKILL.md packages—usable whenever a solo builder needs to codify a repeatable workflow before relying on ad-hoc prompts.
Install
npx skills add https://github.com/himself65/finance-skills --skill skill-creatorWhat is this skill?
- Full skill lifecycle: plan triggers, write SKILL.md, defer detail to reference files, and gate steps with explicit exits
- Eval and benchmark flows with variance analysis for measuring skill quality over iterations
- Rubric-driven scoring when a skill is not working well or needs optimization
- Philosophy: precise triggers and defaults over bloated prose
- Core rule: stay dynamic in content generation, never hard-code one-off examples into the skill artifact
Adoption & trust: 573 installs on skills.sh; 2.7k GitHub stars; 2/3 security scanners passed (skills.sh audits).
What problem does it solve?
You repeat the same multi-step agent instructions in chat and nothing reusable, testable, or discoverable exists in your skills folder.
Who is it for?
Builders who already ship with agent clients and want durable skills with evals when quality regressions would hurt their workflow.
Skip if: One-off tasks with no repetition, or users who only need a single MCP server connection without a SKILL.md procedural package.
When should I use this skill?
User wants to create a skill from scratch, update or optimize a skill, run evals, benchmark performance, or turn a repeatable manual workflow into a skill; triggers include "create a skill", "make a new skill", "evaluate
What do I get? / Deliverables
You ship a scored, trigger-rich skill with reference files and optional eval evidence, ready to install and invoke by name in your agent client.
- SKILL.md with triggers and steps
- Optional reference files and eval artifacts
- Quality rubric scores or benchmark notes
Recommended Skills
Journey fit
Useful at every journey phase - explore requirements and options before committing to a direction.
Where it fits
Capture a competitive-research ritual as a skill with triggers so every new idea starts with the same evidence template.
Turn your MVP scoping checklist into a skill with exit gates before you commit to a full build.
Scaffold SKILL.md plus reference files for a finance or API workflow your repo will invoke by name.
Package a code-review rubric skill and run evals so severity buckets stay consistent across PRs.
Encode a repeatable content or lifecycle playbook once you prove the manual process in chat.
How it compares
Skill package authoring and quality loop—not an MCP integration skill and not a single-phase task script.
Common Questions / FAQ
Who is skill-creator for?
Solo and indie developers using Claude Code, Cursor, or Codex who want to author, tune, or benchmark their own agent skills instead of pasting long prompts each session.
When should I use skill-creator?
Use it in Idea when capturing a repeatable research ritual, in Validate when scoping gets a checklist skill, in Build when adding agent-tooling, in Ship for review or test workflows, and anytime you say create, improve, evaluate, or benchmark a skill.
Is skill-creator safe to install?
It guides file creation and eval runs on your machine—review the Security Audits panel on this page and avoid embedding live secrets or destructive shell defaults into generated skills.
SKILL.md
READMESKILL.md - Skill Creator
# Skill Creator Create, evaluate, and iterate on high-quality agent skills. This skill guides the entire lifecycle: planning what the skill should do, writing SKILL.md and reference files, scoring quality against a rubric, and iterating until the skill meets production standards. **Philosophy:** A great skill is not a long skill. It is a *precise* skill: exhaustive triggers, explicit defaults, clear steps with exit gates, deferred complexity via reference files, and a structured output template. **Core rule — always dynamic, never static:** Skills MUST detect what tools, libraries, and auth are available at runtime and adapt their behavior accordingly. Never hardcode a single method. Always provide a detection flow with a decision tree and fallback paths. See `references/dynamic-calling.md` for the complete pattern catalog. --- ## Step 1: Understand What the User Wants Classify the request into one of these modes: | User Intent | Mode | Jump To | |---|---|---| | Create a brand-new skill | **Create** | Step 2 | | Improve / fix an existing skill | **Improve** | Step 6 | | Evaluate / score a skill's quality | **Evaluate** | Step 7 | If ambiguous, ask: "Do you want to create a new skill, improve an existing one, or evaluate one?" ### Gather Requirements (for Create mode) Before writing anything, answer these questions (ask the user if unclear): | Question | Why it matters | |---|---| | What task does the skill automate? | Defines the core workflow | | Who is the target user? | Determines complexity and terminology level | | What tools/APIs/CLIs does it use? | Determines dependencies and platform restrictions | | What does the user provide as input? | Defines parameters and defaults | | What should the output look like? | Defines the response template | | Does it need API keys or credentials? | Determines `required_environment_variables` | | Should it work on Claude.ai or only CLI? | Determines platform field and dynamic commands | --- ## Step 2: Plan the Skill Architecture Before writing SKILL.md, plan the structure. Read `references/architecture-patterns.md` for detailed guidance on each pattern. ### Choose a Structural Pattern | Pattern | When to use | Steps | Example | |---|---|---|---| | **Linear** | Single workflow, no branching | 5-7 | earnings-preview, etf-premium | | **Router** | Multiple sub-tasks under one umbrella | 3 + sub-skills | stock-correlation (4 sub-skills) | | **Methodology** | Complex domain framework with sequential gates | 7-9 | sepa-strategy (9-step trading methodology) | | **Widget** | Generates interactive UI output | 4-5 | options-payoff (extract + compute + render) | | **API Wrapper** | Wraps an external API with many endpoints | 3-5 + heavy references | funda-data (5 steps, 8 reference files) | ### Plan the Step Outline Write out the step names before writing content. Every skill should have: 1. **Detection flow** (Step 1) -- dynamically detect available tools, auth state, and runtime environment; build a decision tree for which method to use