Skill Creator

Skill Creator

Create and update effective Codex skills. For Codex Lab or another compatible host, verify its current capabilities before relying on host-specific behavior. Apply task scope and authorization; preserve the user's existing authorization and intentional quality and approval policies. Install this maintained override with the catalog's shared ../references/ directory; copying this skill folder alone does not include those dependencies.

About Skills

Skills are modular, self-contained folders that extend Codex's capabilities by providing specialized knowledge, workflows, and tools. Think of them as "onboarding guides" for specific domains or tasks—they transform Codex from a general-purpose agent into a specialized agent equipped with procedural knowledge that no model can fully possess.

What Skills Provide

1. Specialized workflows - Multi-step procedures for specific domains
2. Tool integrations - Instructions for working with specific file formats or APIs
3. Domain expertise - Company-specific knowledge, schemas, business logic
4. Bundled resources - Scripts, references, and assets for complex and repetitive tasks

Core Principles

Concise is Key

The context window is a public good. Skills share the context window with everything else Codex needs: system prompt, conversation history, other Skills' metadata, and the actual user request.

Default assumption: Codex is already very smart. Only add context Codex doesn't already have. Challenge each piece of information: "Does Codex really need this explanation?" and "Does this paragraph justify its token cost?"

Prefer concise examples over verbose explanations.

Set Appropriate Degrees of Freedom

Match the level of specificity to the task's fragility and variability:

High freedom (text-based instructions): Use when multiple approaches are valid, decisions depend on context, or heuristics guide the approach.

Medium freedom (pseudocode or scripts with parameters): Use when a preferred pattern exists, some variation is acceptable, or configuration affects behavior.

Low freedom (specific scripts, few parameters): Use when operations are fragile and error-prone, consistency is critical, or a specific sequence must be followed.

Think of Codex as exploring a path: a narrow bridge with cliffs needs specific guardrails (low freedom), while an open field allows many routes (high freedom).

Write Outcome-First Contracts

Describe the destination and completion bar before prescribing process. A substantial skill should make five things easy to find:

1. Goal - the user-visible outcome the skill owns.
2. Success evidence - what must be true before the task is complete.
3. Autonomy boundary - which safe local actions are allowed and which destructive, external, costly, or scope-expanding actions require approval.
4. Tool routing - the primary helper or decision rule for choosing tools.
5. Final result contract - the status, evidence, gaps, and next action the user should receive.

Keep hard safety, permission, evidence, and output constraints. Remove repeated rules, examples, or process narration that do not change behavior. Use absolute language only for true invariants; use decision rules for judgment calls.

Protect Validation Integrity

You may use subagents during iteration to validate whether a skill works on realistic tasks or whether a suspected problem is real. This is most useful when you want an independent pass on the skill's behavior, outputs, or failure modes after a revision. Only do this when it is possible to start new subagents.

When a skill change alters approval, safety, or destructive behavior, read reviews by another model before merging it: it says when to ask a different provider's model for a review and how to weigh the findings without adding cases nobody can produce.

When using subagents for validation, treat that as an evaluation surface. The goal is to learn whether the skill generalizes, not whether another agent can reconstruct the answer from leaked context.

Prefer raw artifacts such as example prompts, outputs, diffs, logs, or traces. Give the minimum task-local context needed to perform the validation. Avoid passing the intended answer, suspected bug, intended fix, or your prior conclusions unless the validation explicitly requires them.

For changes to routing, command policy, safety boundaries, or GitHub/repo workflow semantics, normally cover at least three representative prompts: the intended trigger and success path, an adjacent request that should route elsewhere, and a boundary case, using the evidence type selected below. Include a negative or ambiguity case when practical. Keep checks focused on observable behavior.

Read validation guidance when selecting evidence for a behavior-sensitive change. Use the current Codex or Codex Lab execution path when testing that host. A direct local-model response or a source review has a different scope from an agent execution test; report which one was performed. Every Code is retired; its harness is not a supported validation path. Retained fixtures and artifact readers are historical material.

Read model-aware authoring when adapting a skill for a named model or investigating a model-specific behavior change. Use openai-docs for current OpenAI model guidance. Keep shared policies in their existing owners; model-specific experiments do not become permanent defaults.

Anatomy of a Skill

Every skill consists of a required SKILL.md file and optional bundled resources:

skill-name/
├── SKILL.md (required)
│   ├── YAML frontmatter metadata (required)
│   │   ├── name: (required)
│   │   ├── description: (required)
│   │   ├── metadata.short-description: (optional)
│   │   └── catalog extensions: e.g. policy.command_policies (optional)
│   └── Markdown instructions (required)
├── agents/ (recommended)
│   └── openai.yaml - UI metadata, dependencies, and invocation policy
└── Bundled Resources (optional)
    ├── scripts/          - Executable code (Python/Bash/etc.)
    ├── references/       - Documentation intended to be loaded into context as needed
    └── assets/           - Files used in output (templates, icons, fonts, etc.)

SKILL.md (required)

Every SKILL.md consists of:

• Frontmatter (YAML): Contains name and description. description is the full model-visible trigger/routing text. Optional metadata.short-description provides compact listing text.
• Repository metadata (YAML): This catalog also supports resources, commands, workflow_defaults, and command-policy declarations for its tooling. These extensions are not Codex runtime enforcement controls.
• Body (Markdown): Instructions and guidance for using the skill. Only loaded AFTER the skill triggers (if at all).
• agents/openai.yaml: Codex UI metadata, tool dependencies, and invocation policy. Put policy.allow_implicit_invocation: false here for explicit-only invocation.

Structured resources and commands

Preserve metadata consumed by this repository's tooling. Add these extensions only when the target catalog uses them; a portable Codex skill does not need them. Keep essential routing and constraints in the body. Read skill design details when adding or changing structured metadata.

Command policies

When this catalog owns a fragile or preferred command workflow, preserve its machine-readable mapping in policy.command_policies as well as the essential helper-routing instruction in prose. Codex reading that metadata is not proof that a command will be intercepted.

Use references/command-policy-contract.md as the source of truth for the frontmatter/runtime boundary, matcher precedence, path resolution, and execution-evidence limits. Keep one canonical owner for each raw command path and use the narrowest matcher that represents the workflow. Command policies are portable metadata, not a runtime enforcement guarantee by themselves.

Agents metadata (recommended)

• UI-facing metadata for skill lists and chips
• Read references/openai_yaml.md before generating values and follow its descriptions and constraints
• Create: human-facing display_name, short_description, and default_prompt by reading the skill
• Generate deterministically by passing the values as --interface key=value to scripts/generate_openai_yaml.py or scripts/init_skill.py
• On updates: check that agents/openai.yaml still matches the skill and edit the intended fields in place. Preserve existing policy, dependencies, and unrelated interface fields. The bundled generator writes an interface-only file, so use it for initial generation, not an existing metadata update.
• Only include other optional interface fields (icons, brand color) if explicitly provided
• See references/openai_yaml.md for field definitions and examples

Bundled Resources (optional)

Use scripts for deterministic or repeatedly rewritten work, references for details loaded only when needed, and assets for files copied or modified in the final output. Python helpers require PEP 723 metadata and should run with uv run; shell helpers must be documented and invoked as shell helpers. Prefer stdin or body files for fragile multiline payloads. Keep information in one place rather than duplicating it between SKILL.md and references.

Read references/skill-design-details.md for resource patterns, helper invocation rules, and progressive-disclosure examples. Skills that generate reviews, handoffs, issue or PR comments, readiness reports, or final summaries should point to ../references/every-code-formatting.md instead of copying its rules.

What to Not Include in a Skill

A skill should only contain essential files that directly support its functionality. Do NOT create extraneous documentation or auxiliary files, including:

• README.md
• INSTALLATION_GUIDE.md
• QUICK_REFERENCE.md
• CHANGELOG.md
• etc.

The skill should only contain the information needed for an AI agent to do the job at hand. It should not contain auxiliary context about the process that went into creating it, setup and testing procedures, user-facing documentation, etc. Creating additional documentation files just adds clutter and confusion.

Progressive Disclosure Design Principle

Skills use a three-level loading system to manage context efficiently:

1. Metadata (name + description) - Always in context (~100 words)
2. SKILL.md body - When skill triggers (<5k words)
3. Bundled resources - As needed by Codex (Unlimited because scripts can be executed without reading into context window)

Progressive Disclosure Patterns

Keep the body to essential routing, judgment, safety, and workflow instructions and under 500 lines. Move variant-specific details, examples, configuration, schemas, and exhaustive command documentation to one-level-deep references. Name each reference from SKILL.md and say when to read it. Add a table of contents to longer reference files. See references/skill-design-details.md for worked patterns.

Skill Creation Process

Skill creation involves these steps:

1. Understand the skill with concrete examples
2. Plan reusable skill contents (scripts, references, assets)
3. Initialize the skill (run init_skill.py)
4. Edit the skill (implement resources and write SKILL.md)
5. Validate the skill (run quick_validate.py)
6. Iterate based on real usage and forward-test complex skills.

Follow these steps in order, skipping only if there is a clear reason why they are not applicable.

Skill Naming

• Use lowercase letters, digits, and hyphens only; normalize user-provided titles to hyphen-case (e.g., "Plan Mode" -> plan-mode).
• When generating names, generate a name under 64 characters (letters, digits, hyphens).
• Prefer short, verb-led phrases that describe the action.
• Namespace by tool when it improves clarity or triggering (e.g., gh-address-comments, linear-address-issue).
• Name the skill folder exactly after the skill name.

Step 1: Understanding the Skill with Concrete Examples

Reuse usage patterns and examples already clear from the request, conversation, or existing skill. Ask only for missing information that materially affects the skill's purpose, audience, or behavior.

To create an effective skill, clearly understand concrete examples of how the skill will be used. This understanding can come from either direct user examples or generated examples that are validated with user feedback.

For example, when building an image-editor skill, relevant questions include:

• "What functionality should the image-editor skill support? Editing, rotating, anything else?"
• "Can you give some examples of how this skill would be used?"
• "I can imagine users asking for things like 'Remove the red-eye from this image' or 'Rotate this image'. Are there other ways you imagine this skill being used?"
• "What would a user say that should trigger this skill?"

To avoid overwhelming users, avoid asking too many questions in a single message. Start with the most important questions and follow up as needed for better effectiveness.

Conclude this step when there is a clear sense of the functionality the skill should support.

Step 2: Planning the Reusable Skill Contents

To turn concrete examples into an effective skill, analyze each example by:

1. Considering how to execute on the example from scratch
2. Identifying what scripts, references, and assets would be helpful when executing these workflows repeatedly

Example: When building a pdf-editor skill to handle queries like "Help me rotate this PDF," the analysis shows:

1. Rotating a PDF requires re-writing the same code each time
2. A scripts/rotate_pdf.py script would be helpful to store in the skill

Example: When designing a frontend-webapp-builder skill for queries like "Build me a todo app" or "Build me a dashboard to track my steps," the analysis shows:

1. Writing a frontend webapp requires the same boilerplate HTML/React each time
2. An assets/hello-world/ template containing the boilerplate HTML/React project files would be helpful to store in the skill

Example: When building a big-query skill to handle queries like "How many users have logged in today?" the analysis shows:

1. Querying BigQuery requires re-discovering the table schemas and relationships each time
2. A references/schema.md file documenting the table schemas would be helpful to store in the skill

To establish the skill's contents, analyze each concrete example to create a list of the reusable resources to include: scripts, references, and assets.

Step 3: Initializing the Skill

At this point, it is time to actually create the skill.

Skip this step only if the skill being developed already exists. In this case, continue to the next step.

Use an explicit location or the established maintained source first. In a skills repository, follow its branch/worktree and runtime-checkout instructions; do not scaffold into the installed runtime or generated .system/plugin cache. For a new Codex skill outside an established catalog, use the intended scope: .agents/skills in the repository for project skills, ~/.agents/skills for personal skills, or the owning plugin's source directory for plugin skills. These follow the current Codex discovery documentation. Ask about location only when the intended scope remains materially ambiguous.

Resolve existing symlinks and host configuration before choosing a destination. CODE_HOME, $CODEX_HOME/skills, and .code/skills may describe compatibility or existing runtime layouts; they are not universal Codex authoring defaults. For another host, verify its current discovery rules in that host's documentation or maintained source before using a different layout.

When creating a new skill from scratch, always run the init_skill.py script. The script conveniently generates a new template skill directory that automatically includes everything a skill requires, making the skill creation process much more efficient and reliable.

Commands below are relative to this skill-creator directory; from another directory, use the script's absolute path. Use uv for the bundled Python helpers and their declared dependencies. If setup is needed, use python-uv-workflow.

Usage:

uv run scripts/init_skill.py <skill-name> --path <maintained-skill-root> [--resources scripts,references,assets] [--examples]

Examples:

uv run scripts/init_skill.py my-skill --path <target-repo-root>/.agents/skills
uv run scripts/init_skill.py my-skill --path <target-repo-root>/.agents/skills --resources scripts,references
uv run scripts/init_skill.py my-skill --path <maintained-skill-root> --resources scripts --examples

The script:

• Creates the skill directory at the specified path
• Generates a SKILL.md template with proper frontmatter and TODO placeholders
• Creates agents/openai.yaml using agent-generated display_name, short_description, and default_prompt passed via --interface key=value
• Optionally creates resource directories based on --resources
• Optionally adds example files when --examples is set

After initialization, customize the SKILL.md and add resources as needed. If you used --examples, replace or delete placeholder files.

For initial UI metadata, derive display_name, short_description, and default_prompt from the skill and pass them as --interface key=value to init_skill.py. If agents/openai.yaml does not yet exist, it can also be generated with:

uv run scripts/generate_openai_yaml.py <path/to/skill-folder> --interface key=value

For an existing metadata file, edit only the intended fields in place; do not use that generator to update it. Preserve policy, dependencies, and unrelated interface fields.

Only include other optional interface fields when the user explicitly provides them. For full field descriptions and examples, see references/openai_yaml.md.

Step 4: Edit the Skill

When editing the (newly-generated or existing) skill, remember that the skill is being created for another coding agent instance to use. Include information that would be beneficial and non-obvious to that agent. Consider what procedural knowledge, domain-specific details, or reusable assets would help another agent execute these tasks more effectively.

After substantial revisions, or if the skill is particularly tricky, you should use subagents to forward-test the skill on realistic tasks or artifacts. When doing so, pass the artifact under validation rather than your diagnosis of what is wrong, and keep the prompt generic enough that success depends on transferable reasoning rather than hidden ground truth.

Start with Reusable Skill Contents

To begin implementation, start with the reusable resources identified above: scripts/, references/, and assets/ files. Note that this step may require user input. For example, when implementing a brand-guidelines skill, the user may need to provide brand assets or templates to store in assets/, or documentation to store in references/.

Added scripts must be tested by actually running them to ensure there are no bugs and that the output matches what is expected. If there are many similar scripts, only a representative sample needs to be tested to ensure confidence that they all work while balancing time to completion.

If you used --examples, delete any placeholder files that are not needed for the skill. Only create resource directories that are actually required.

Update SKILL.md

Writing Guidelines: Always use imperative/infinitive form.

Frontmatter

Write Codex YAML frontmatter with name and description:

• name: The skill name
• description: This is the primary triggering mechanism for your skill, and helps the agent understand when to use the skill.
  • Include both what the Skill does and specific triggers/contexts for when to use it.
  • Include all "when to use" information here - Not in the body. The body is only loaded after triggering, so "When to Use This Skill" sections in the body are not helpful to the agent.
  • Example description for a docx skill: "Comprehensive document creation, editing, and analysis with support for tracked changes, comments, formatting preservation, and text extraction. Use when Codex needs to work with professional documents (.docx files) for: (1) Creating new documents, (2) Modifying or editing content, (3) Working with tracked changes, (4) Adding comments, or any other document tasks"
• metadata.short-description: Optional compact human-facing summary for UI/listing surfaces. Keep the full routing and trigger detail in description.

Example:

---
name: example-skill
description: Full model-visible trigger/routing description. Use when ...
metadata:
  short-description: Compact human-facing summary
---

For a Codex skill requiring explicit invocation, add this to agents/openai.yaml, preserving any existing interface and dependencies:

policy:
  allow_implicit_invocation: false

Do not rely on a same-named SKILL.md frontmatter field for Codex invocation control. Preserve such fields only for a catalog or host that consumes them, and document that compatibility scope. Use runtime config for installation-only selection. Consult agents metadata for supported fields and skill design details for this repository's additional frontmatter metadata.

Body

Write instructions for using the skill and its bundled resources.

Step 5: Validate the Skill

Once development of the skill is complete, validate the skill folder to catch basic issues early:

uv run scripts/quick_validate.py <path/to/skill-folder>

The local validator checks frontmatter, naming, and this catalog's metadata contracts. It does not prove target-host discovery or model behavior. Complete the owning repository's required checks; select additional evidence using validation guidance. If a check fails, fix the issue and rerun the affected check.

Step 6: Iterate

After testing the skill, forward-test substantial or fragile changes and use fresh user feedback to identify measured failures.

Forward-testing and iteration workflow:

1. Reuse an applicable baseline on representative tasks, or record one when making a behavior or performance comparison.
2. Identify a concrete struggle, inefficiency, or contract failure.
3. Change one instruction group, resource, or tool route at a time.
4. Run affected cases and compare observable outcomes with any applicable baseline; reuse current evidence for unchanged scope and inputs.
5. Keep supported improvements and report unmeasured effects or unavailable execution evidence without claiming a demonstrated speedup.

Read references/forward-testing.md only when planning or running a subagent forward-test for a tricky skill revision.