gpt-image-skill

Generate or edit images using OpenAI GPT Image API (gpt-image-2, gpt-image-1, etc). Use ONLY when the user explicitly names OpenAI or GPT as the provider: "gpt image", "openai image", "generate image with openai", "用 openai 画图", "用 GPT 生成图片". For generic image requests without a provider, use nanobanana-skill instead. Do NOT use for diagrams (架构图/流程图) — draw those with Mermaid or code.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

82%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 highly actionable, executable body that documents the bundled CLI exhaustively with copy-paste commands, defaults, and error recovery. Its weaknesses are minor: duplicated option listings between the Instructions and Available Options sections, and no explicit output-validation checkpoint in the workflow.

Dimension	Reasoning	Score
Conciseness	The body is almost entirely commands, option tables, and examples with no explanation of concepts Claude already knows; it does not earn a 5 because the options are listed twice ("Ask the user for: ... Desired size ... Model preference ... Quality ... Number of images" and again under "Available Options") and the eight examples overlap in flag usage. Not a 3 — there is no padded prose, only minor duplication that could be trimmed.	4 / 5
Actionability	Fully executable copy-paste commands throughout: "python3 ${CLAUDE_SKILL_DIR}/gpt_image.py --prompt ... --output ..." plus edit subcommand with "--input image1.png image2.png", install command ("python3 -m pip install -r ${CLAUDE_SKILL_DIR}/requirements.txt"), and every flag documented with defaults. Anchor 5: covers all common cases copy-paste ready.	5 / 5
Workflow Clarity	Clear sequences for both generation ("1. Ask the user ... 2. Run the script ... 3. Show the user the saved image path") and editing, with an Error Handling section giving recovery steps ("Check that OPENAI_API_KEY is exported", "verify OPENAI_API_BASE is correct"). Not a 5 because there is no explicit validate-before-proceeding checkpoint; the skill is not destructive/batch so no cap applies.	4 / 5
Progressive Disclosure	No bundle files are present (no references/, scripts/, or assets/); the body is a single ~180-line document with well-organized sections (Requirements, Instructions, Available Options, Examples, Error Handling, Best Practices) and the script dependency clearly signaled via ${CLAUDE_SKILL_DIR}/gpt_image.py. Good structure earns a 4; not a 5 because all option reference material is inlined rather than split out, and it exceeds the under-50-line simple-skill case.	4 / 5
	Total	17 / 20 Passed

Description

90%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.

A strong description with explicit bilingual trigger phrases, clear provider-boundary guidance, and explicit disambiguation from competing skills. Its only weakness is capability specificity limited to two actions (generate, edit) with model names, rather than a broader inventory of concrete capabilities.

Dimension	Reasoning	Score
Specificity	"Generate or edit images using OpenAI GPT Image API (gpt-image-2, gpt-image-1, etc)" names the domain with exactly two concrete actions (generate, edit) plus model enumeration, but offers no broader capability coverage. This matches the anchor for 1-2 concrete actions without comprehensiveness; a 4 would require several distinct specific actions.	3 / 5
Completeness	Both questions are explicitly answered: the 'what' ("Generate or edit images using OpenAI GPT Image API") and the 'when' ("Use ONLY when the user explicitly names OpenAI or GPT as the provider") with concrete trigger phrases plus negative guidance. Anchor 5; a 4 would have a less explicit 'when'.	5 / 5
Trigger Term Quality	Explicit natural trigger phrases are quoted in two languages: "gpt image", "openai image", "generate image with openai", "用 openai 画图", "用 GPT 生成图片" — comprehensive synonym coverage users would naturally say. Anchor 5 applies; nothing common is missing.	5 / 5
Distinctiveness Conflict Risk	The description carves a clear niche — it names the exclusive trigger condition (user must name OpenAI/GPT), redirects generic requests to "nanobanana-skill instead", and excludes diagrams ("draw those with Mermaid or code"). Minimal conflict risk; anchor 5.	5 / 5
	Total	18 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning

	Total	15 / 16 Passed

Repository: feiskyer/claude-code-settings
Path: skills/gpt-image-skill/SKILL.md
Commit: 95dab59

Reviewed: 8 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.