gpt-image

Use this skill whenever a user asks to generate, create, draw, render, or edit images with GPT Image 2 / gpt-image-2, text-to-image, reference-image editing, inpainting, posters, typography, Chinese text, UI mockups, diagrams, or gallery prompts. Analyze the user's prompt, search the bundled Reference Gallery/craft files for matching design patterns, confer on direction when useful, then call the packaged `gpt-image` CLI or bundled `scripts/generate.py`. Do not write new image-generation code unless explicitly asked to modify this repo.

Quality

89%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Quality

Content

78%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-organized runbook that pairs concrete CLI commands and tables with a clear operating loop and genuine one-level-deep reference disclosure. Its weakest spots are minor redundancy in the CLI resolution chain and the lack of an explicit error-retry feedback loop in verification.

Suggestions

Consolidate the CLI fallback chain into one place (CLI resolution) and have the Operating loop and generate.py note reference it, to remove the triplicate restatement.

Add one fully spelled-out end-to-end example invocation (real prompt text + a representative flag combination) to lift actionability from mostly-executable to copy-paste-ready.

Extend the Verification section with an explicit retry loop (e.g., on exit code 1 or missing-key errors: diagnose from stderr, fix the flag/path or key, re-run) to close the workflow-clarity feedback-loop gap.

Dimension	Reasoning	Score
Conciseness	The body is dense and mostly tables/rules with no padding of concepts Claude already knows, but the CLI fallback chain is restated in the Operating loop, the CLI resolution section, and the generate.py launcher description, leaving minor redundancy that could be trimmed.	4 / 5
Actionability	Provides copy-paste-ready CLI invocations (`gpt-image -p "PROMPT" [-f OUT] [-i REF...] [-m MASK]`), a complete flags table with concrete values, and an endpoint routing table; falls short of a 5 only because example commands use placeholder tokens rather than a fully spelled-out end-to-end invocation.	4 / 5
Workflow Clarity	An explicit 8-step Operating loop plus a Verification section with before/after checks, exit-code semantics (0/1/2), and path-existence checks gives a clear sequence with most checkpoints; a full 5 would require an explicit error-recovery/retry loop rather than just surfacing stderr.	4 / 5
Progressive Disclosure	The body is an overview pointing one level deep to real, well-signaled reference files (gallery.md routing index, gallery-*.md categories, craft.md, openai-cookbook.md) — all verified present in ./references/ — with an explicit 'load the smallest useful slice; never load all category files by default' policy.	5 / 5
	Total	17 / 20 Passed

Description

100%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is a strong, third-person trigger statement that concretely names what the skill does and when to use it, with rich natural synonyms and a model-specific niche that minimizes conflict risk. It cleanly matches the top anchor on every dimension.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'generate, create, draw, render, or edit images', 'search the bundled Reference Gallery/craft files for matching design patterns', 'call the packaged `gpt-image` CLI or bundled `scripts/generate.py`', 'Do not write new image-generation code' — with comprehensive coverage of the skill's behavior.	5 / 5
Completeness	Explicitly answers 'when' via 'Use this skill whenever a user asks to generate, create, draw, render, or edit images...' and 'what' via 'Analyze the user's prompt, search... confer... then call the packaged `gpt-image` CLI', satisfying both halves with concrete trigger phrases.	5 / 5
Trigger Term Quality	Comprehensive natural-term coverage with synonyms ('generate, create, draw, render, or edit'), the specific model name 'GPT Image 2 / gpt-image-2', and concrete modality triggers like 'text-to-image, reference-image editing, inpainting, posters, typography, Chinese text, UI mockups, diagrams'.	5 / 5
Distinctiveness Conflict Risk	The trigger is tied to a specific named model ('GPT Image 2 / gpt-image-2') and a distinct set of imaging modalities, giving it a clear niche with minimal overlap risk against other skills.	5 / 5
	Total	20 / 20 Passed

Validation

81%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
referenced_paths_exist	Referenced path issues: 2 missing	Warning

	Total	13 / 16 Passed

Repository: wuyoscar/GPT-Image2-Skill
Path: skills/gpt-image/SKILL.md
Commit: 068dd9e

Reviewed: 18 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.