What does this skill do, and when should you use it?
Skill Creator is a workflow guide for developers building skills for Codex. It explains how to plan skill contents, structure SKILL.md, and add optional scripts, references, and assets while managing context through progressive disclosure. It also covers naming, initialization, editing, validation, and iteration. It is a good fit for Codex skill authors, but its instructions are explicitly centered on Codex’s skill structure and tooling.
Guides users through analyzing concrete use cases, planning scripts, references, and assets, initializing a skill with init_skill.py, writing SKILL.md with name and description metadata, generating or updating agents/openai.yaml, and validating the result with quick_validate.py. It provides written rules for skill anatomy, trigger descriptions, resource organization, naming, and progressive disclosure.
- A developer creating a new Codex skill for a specialized domain can follow the end-to-end authoring workflow.
- A team encapsulating repetitive or reliability-sensitive code can use the guidance to decide when to bundle scripts.
- A developer documenting schemas, policies, or API knowledge can organize on-demand context with references.
- An owner maintaining an existing skill can revise SKILL.md or bundled resources based on real usage and validate the changes.
How do you install this skill?
- Before use, confirm script scope, write targets, and overwrite behavior, and keep backups; the documentation does not provide a rollback procedure.
- License, versioning, maintenance ownership, and update sources are not clearly established within the assessed skill path and should be verified before distribution.
- The default prompt does not explicitly include $skill-creator as required by the referenced convention, which may reduce invocation consistency.
- No scripts were executed and no real outputs were independently tested; this is not evidence of runtime reliability or result quality.
- Shell / CLI
- Local filesystem
The skill is located at skills/.system/skill-creator/SKILL.md. The README states that .system skills are automatically installed with the latest version of Codex; the supplied source does not provide a separate manual installation command. It says an individual skill’s license should be checked in its directory’s LICENSE.txt, but that file’s contents are not supplied here.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/openai/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/.system/skill-creator" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Create a skill for PDF processing
- Update this skill to support a new workflow.
In Codex, request a new or updated skill, for example, “Create a skill for PDF processing” or “Update this skill to support a new workflow.” The guide specifies an order of understanding examples, planning reusable contents, running scripts/init_skill.py, editing SKILL.md, running scripts/quick_validate.py, and iterating from real usage; exact script arguments should be confirmed from the local scripts or environment.
What are this skill's strengths and limitations?
- Covers the workflow from requirements discovery through validation and iteration.
- Clearly separates the roles of SKILL.md, scripts, references, and output assets.
- Emphasizes concise instructions, progressive disclosure, and task-appropriate degrees of freedom.
- Defines naming, frontmatter, and validation expectations.
- The guidance is written for creating Codex skills and may require adaptation for other Agent Skills clients.
- The supplied material does not include the referenced helper-script implementations, so their platform behavior cannot be independently assessed.
- No test suite, platform coverage, or individual skill license text is provided.
- The repository README marks the repository as deprecated and points readers to the OpenAI Plugins repository and newer plugin-building guidance.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| OpenAI Skill Creator Guide this page ✓ OpenAI · Official | 55 · Use with care | ★ 28k | 4mo ago | — |
| Skill Builder Guide ✓ Anthropic · Official | 52 · Use with care | ★ 40k | 1mo ago | Apache-2.0 |
| Codex Sample Skill Creator Guide ✓ OpenAI · Official | 46 · Use with care | ★ 129k | 11d ago | Apache-2.0 |
| Claude Code Plugin Skill Development ✓ Anthropic · Official | 51 · Use with care | ★ 38k | 11d ago | Apache-2.0 |
| Workflow Skill Builder ✓ Microsoft · Official | 50 · Use with care | ★ 194k | 11d ago | MIT |
How did FollowSkills review this skill?
The skill only guides skill creation, configuration generation, and local validation; no malware, credential theft, or covert exfiltration is evident. However, user confirmation, permission boundaries, data flow, rollback, sensitive-data handling, and license details are not explicit, so points are deducted.
The creation, editing, and validation workflow is reasonably coherent, and the scripts appear aligned on the main path. This was a static review with no execution, and there is no committed test suite, CI coverage, or comprehensive abnormal-input diagnosis; the default prompt also conflicts with the cited prompt convention, so the score is conservative.
The name, trigger description, audience, and main scenarios are clear for creating or updating skills. Non-fit boundaries, false-trigger constraints, Chinese-language guidance, and mainland-China environment fit are insufficiently evidenced, so the score is reduced.
The documentation has clear progressive-disclosure and resource-organization guidance, plus naming and validation instructions. License, versioning, changelog, maintenance ownership, update path, troubleshooting, and FAQ coverage are incomplete, so points are deducted.
The skill provides a template, workflow, and helper scripts that can directly support skill creation. However, final output still depends on substantial author editing, representative results were not execution-verified, and comparative benefit over manual work is not evidenced; the static cap therefore applies.
The supplied files and script logic are auditable and provide limited reproducibility. There are no execution records, third-party reproduction results, committed tests, or CI evidence, so the score remains below the static maximum.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →