---
title: "Anthropic Skill Creator: review, setup & limitations · Undominated.ai"
canonical: https://undominated.ai/skills/anthropics-skill-creator/
description: "Develops skills from realistic task fixtures and baseline comparisons, with separate trigger-description experiments and human output review."
---

# Anthropic Skill Creator: review, setup & limitations · Undominated.ai

> Develops skills from realistic task fixtures and baseline comparisons, with separate trigger-description experiments and human output review.

[← Explore all skills](/skills/)

AGENT DEVELOPMENT / Anthropic

# Anthropic Skill Creator

Develops skills from realistic task fixtures and baseline comparisons, with separate trigger-description experiments and human output review.

 [See setup guidance ↓](#setup)[Original source ↗](https://github.com/anthropics/skills/tree/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator)

SOURCE REVIEW

 Reviewed 2026-10-15
 Evidence 9 linked sources
 Publisher Anthropic
 Licence [Apache-2.0 ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/LICENSE.txt)
 Revision 34040c9c5685
 [Read what was—and wasn’t—checked ↓](#review)

“Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.”

 [Anthropic · upstream description ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/SKILL.md) Our analysis follows below.

01 / THE REASONING

## Why this made the selection.

 - Separates activation experiments from task-output quality, preserving both a baseline and human feedback.
- Provides concrete task, grading and benchmark schemas and an inspectable trigger evaluator rather than treating a well-formed SKILL.md as proof of usefulness.

### A good fit for

 - Turning a repeated workflow, example files and acceptance criteria into a skill plus evaluation fixtures and a human-review packet.
- Comparing a revised skill against its original, with separate positive/negative activation queries and retained iteration history.

### Weigh up before choosing

 - Full comparisons require supported subagents and Claude CLI; reduced Claude.ai paths explicitly omit baseline benchmarking. A source review supplies neither benchmark results nor installation proof.
- run_eval.py detects the first matching Skill/Read event for a temporary command stub; it does not measure completion of the real skill body. A timeout or worker exception becomes false, which can make a negative query look like a correct non-trigger.
- run_loop.py selects the best description using repeatedly observed held-out test scores. That set is a selection set, not an untouched final estimate of generalization; reserve new queries for final assessment.
- run_eval.py aggregates by query string and run_loop.py separates splits by query-string membership. Duplicate query text can merge repeats or mix train/test attribution; use unique, consistently labeled queries.
- quick_validate.py checks metadata shape and length, not safety or task quality; blank name/description strings can pass its conditional checks. schemas.md calls task checks expectations while SKILL.md calls them assertions.
- CLI trials inherit host permissions and environment, can act on real files/services, and consume model usage. Viewer, packaging, aggregation and remaining agent helpers were not audited in this pass.

02 / THE REVIEW RECORD

## What we actually inspected.

Source review has boundaries. A clear record is more useful than a “safe” badge.

### Material inspected

 - skills/skill-creator/SKILL.md
- skills/skill-creator/references/schemas.md
- skills/skill-creator/scripts/quick_validate.py
- skills/skill-creator/scripts/run_eval.py
- skills/skill-creator/scripts/run_loop.py
- skills/skill-creator/scripts/improve_description.py
- skills/skill-creator/scripts/utils.py
- skills/skill-creator/LICENSE.txt
- README.md and pinned Skills installer README source-format sections

### Our findings

 - Task inputs are a skill directory, realistic prompts, expected outputs, example files and assertions; trigger inputs are query/should_trigger pairs. Task handoffs include iteration outputs, grading.json expectations with text/passed/evidence, timing.json, benchmark.json and human feedback. Trigger output contains trigger_rate, triggers, runs and pass per query.
- The evaluator creates a uniquely named .claude/commands Markdown stub near the current project, invokes claude -p, kills it after detecting the first relevant event or timing out, and removes the stub in finally. This is not an isolated trial workspace.
- The optimizer splits positives/negatives, blinds test fields from description-improvement prompts, and chooses best_description by test_passed. improve_description.py sends the full skill text and failed training queries through Claude CLI and optionally logs those prompts/responses.
- The same pin remains necessary for the catalogue's existing root-discovery receipt; neither that receipt nor install command was changed or upgraded.

### Not established by this review

 - No Claude subprocess, subagent, metadata validator, viewer, installer or evaluation was run.
- No runtime schema compatibility, trigger accuracy, skill effectiveness, packaging safety or dependency-advisory scan. PyYAML, Claude CLI and the Skills installer implementation are not exhaustively audited.
- Not executed: it was not installed or run.

The review applies to the material and revision named here. A newer upstream release can change its behavior.

03 / PUT IT TO WORK

## Add a skill to your workflow.

[Upstream setup instructions ↗](https://github.com/vercel-labs/skills/blob/7407f3893ad4dceab546ac002c3ef806e4000c73/README.md)

DOCUMENTED COMMAND

 npx skills add anthropics/skills --skill skill-creator Copy command ↗

Run in your terminal to choose your AI tools, project or global scope, and installation method. This short command follows the current upstream version; the review covers the revision shown above.

 Optional: install the reviewed revision

Keeps the original source and installer version where recorded, while leaving your AI tools and installation scope for you to choose.

 npx skills add https://github.com/anthropics/skills/tree/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator Copy reviewed-revision command ↗ - Install the complete skill directory with its references and helpers; a copied SKILL.md alone may be incomplete.
- Choose your AI tools and project or global scope in the installer, then review its permissions before using the skill.

### Before you start

 - Node.js/npm for this installer
- An agent that supports the Agent Skills format
- Python/PyYAML for metadata validation
- Claude CLI and supported subagents for full evaluation

### Compatibility

Claude Code · Claude.ai and Cowork with documented workflow limits

### Write skill directories, evaluation artifacts and temporary command files

 - Write skill and evaluation files and temporary .claude/commands files in the discovered project.
- Launch Claude CLI using inherited authentication/environment; improvement calls transmit skill text and example queries to the configured model service.
- Trials can exercise the target skill's own permissions. Optional report generation opens a browser or writes HTML and logs.

### Cost model

The skill is Apache-licensed; repeated agent evaluations consume the configured service allowance or API usage.

04 / FOLLOW THE EVIDENCE

## The source trail.

Our notes are separate from the original resource. Check upstream before adopting a new version.

 - [Skill lifecycle and handoffs ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/SKILL.md) Checked 2026-10-14 https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/SKILL.md Supports: summary, upstreamDescription, whySelected, bestFor, review
- [Task and review schemas ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/references/schemas.md) Checked 2026-10-14 https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/references/schemas.md Supports: review, limitations
- [Trigger evaluator ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/scripts/run_eval.py) Checked 2026-10-14 https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/scripts/run_eval.py Supports: review, limitations, access
- [Description selection loop ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/scripts/run_loop.py) Checked 2026-10-14 https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/scripts/run_loop.py Supports: review, limitations
- [Description improvement subprocess ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/scripts/improve_description.py) Checked 2026-10-14 https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/scripts/improve_description.py Supports: review, access
- [Metadata checker ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/scripts/quick_validate.py) Checked 2026-10-14 https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/scripts/quick_validate.py Supports: limitations, review
- [Skill licence ↗](https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/LICENSE.txt) Checked 2026-10-14 https://github.com/anthropics/skills/blob/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/LICENSE.txt Supports: license
- [Installer source formats ↗](https://github.com/vercel-labs/skills/blob/7407f3893ad4dceab546ac002c3ef806e4000c73/README.md) Checked 2026-10-14 https://github.com/vercel-labs/skills/blob/7407f3893ad4dceab546ac002c3ef806e4000c73/README.md Supports: install
- [2026-10-15 W20 source re-review at the existing pin ↗](https://raw.githubusercontent.com/anthropics/skills/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/SKILL.md) Checked 2026-10-15 https://raw.githubusercontent.com/anthropics/skills/34040c9c568585f6929bedeaad110ad08f079624/skills/skill-creator/SKILL.md Supports: upstreamDescription, review.notTested, review.reviewedAt

KEEP COMPARING

## Other approaches to consider.

Related by category or shared topics. These are alternatives to inspect, not a measured quality order.

 [### Skill Development for Claude Code Plugins ↗ Guides writing a skill that ships inside a Claude Code plugin: directory layout, a third-person trigger description, an imperative SKILL.md body and progressive disclosure into references/, examples/ and scripts/.](/skills/anthropic-skill-development/)[### OpenAI Evaluate Skill ↗ Connects a local skill review to structural checks, budget explanations and an explicitly separate live benchmark workflow.](/skills/openai-evaluate-skill/)[### Command Development for Claude Code ↗ Explains how to write Claude Code slash-command files: frontmatter fields, $ARGUMENTS and positional arguments, @file references, inline Bash and plugin command patterns.](/skills/anthropic-command-development/)

 [AI Tools ↗](/tools/)[Skills ↗](/skills/)[Agents ↗](/agents/)[MCP Servers ↗](/mcp-servers/)[Workflows ↗](/workflows/)

## Continue your investigation

 - [Choose a toolkit](/tools/)
- [Inspect execution hosts](/agents/)
- [Check tool access](/mcp-servers/)
- [Check project dependencies](/check-repo/)
