Skill Creator

88TrustedShips scripts

Automated framework for building, testing, and refining custom AI assistant capabilities.

What it does

Adding this skill transforms your assistant into a development environment for creating new behaviors. It guides you through drafting instructions, defining expected outcomes, and refining the logic through an iterative development cycle. By acting as a partner in the creation process, the assistant helps turn vague workflows or repetitive manual tasks into structured, reliable instructions that the model can execute consistently.

Beyond initial drafting, this skill provides tools for verifying performance. It assists in establishing test cases and generating performance metrics to ensure the resulting logic behaves as intended. Once a skill reaches a functional state, it includes automated refinement utilities to improve triggering accuracy, ensuring the assistant calls upon the correct instructions at the right time.

When it earns its place

Reach for it when…

Best used when you have a specific, repeatable task in mind that requires consistent logic or standardized output formats, especially when you need to verify the accuracy of the assistant's results through formal testing and data-driven iterations.

Look elsewhere when…

Do not use this for quick, one-off tasks that do not require reusable logic. It is also unsuitable for complex software engineering projects or tasks that require integrating external APIs or custom databases, as this toolset is strictly for refining the assistant's internal instruction set.

What to say to use it

You never invoke a skill directly — the assistant picks it up when your request matches what it covers. These are the kinds of things that do that.

Help me turn this workflow into a reusable skill
How can I test if this new instruction set is working correctly
I want to improve how often my assistant triggers this custom skill
Run some benchmarks to see if this skill needs better instructions

Install it

Install for all your projects

Copy the skill directory into your personal skills folder. Claude picks it up on its own the next time your request matches what it does — there is nothing to enable.

bash

git clone --depth 1 --branch main https://github.com/anthropics/skills.git /tmp/skills
mkdir -p ~/.claude/skills
cp -r /tmp/skills/skills/skill-creator ~/.claude/skills/skill-creator

Install for one project only

Put it in the project's own skills folder instead, so it can be committed and shared with everyone working on that repository.

bash

cp -r /tmp/skills/skills/skill-creator .claude/skills/skill-creator

If a skill of the same name exists in both places, the personal one wins.

Installing on Claude.ai or through the API instead? The full guide covers every surface.

Pair it with an MCP server

Skill Creator changes how the work is done. These servers change what Claude can reach while doing it.

How skills and MCP servers differ →

88/100Trusted source

What checks out

Original source. Published in the repository of whoever wrote it, rather than copied out of it into a collection.
Adoption. The host repository has 169,871 stars. This counts the whole repository, not this skill alone.
Official author. Published by the organisation that owns the subject it covers.
Recently maintained. The repository was updated 3 days ago.
Spec-compliant. Frontmatter matches the Agent Skills specification, so it loads in any client implementing it.
Substantive instructions. The body carries enough detail to change how a task is actually performed.

What to know

Licensed. No licence found. Without one you have no explicit permission to reuse or redistribute it.

This score describes where the skill came from — its origin, licence, upkeep and conformance to the spec. It is not a security audit, and it says nothing about whether the skill's instructions are correct. How it is calculated.

What it ships

Executable files: 10 Python files that can run on your machine
Files in total: 18, including reference documents loaded on demand
Context cost: about 8,292 tokens once activated (estimated)

Larger than the 5,000-token body the specification recommends. It still works — it just takes up more of the conversation every time it activates.

Bundled scripts are normal — most skills from established authors include them. We have not run or reviewed this code; read it before installing if that matters to you.

Source

Published in: anthropics/skills
Path in repository: skills/skill-creator/SKILL.md
Licence: Not stated in a standard form

New to skills? Start here →