The body is lean reference material (options, flags, examples) with no padding explaining concepts Claude already knows; every section earns its place. It is not score 2 because there is no unnecessary explanation or looseness to tighten.

Fully executable commands (python3 ${CLAUDE_SKILL_DIR}/gpt_image.py ...) with all flags documented and eight copy-paste-ready examples spanning generation, editing, sizes, backgrounds, and models; not score 2 because guidance is complete and executable rather than pseudocode.

For a simple single-action skill the workflow is unambiguous and sequenced (ask user for params, run script, show saved path) with an Error Handling section for recovery; per the simple-skills note this qualifies for 3, and it is not score 2 because the steps and recovery guidance are explicit.

No bundle files are present, so it is evaluated on body organization: a self-contained file with well-organized sections (Requirements, Instructions, Options, Examples, Error Handling, Best Practices) and no nested reference chains; not score 2 because structure is clear and easy to navigate.

"Generate or edit images using OpenAI GPT Image API" names multiple concrete actions (generate, edit) and specific models, matching the anchor that lists several specific concrete actions; it is not score 2 because the action set is comprehensive rather than partial.

It clearly answers both what ("Generate or edit images using OpenAI GPT Image API") and when ("Use ONLY when the user explicitly names OpenAI or GPT as the provider") with explicit triggers; not score 2 because the when clause is explicit, not merely implied.

Explicit natural trigger phrases are quoted ("gpt image", "openai image", "generate image with openai", "用 openai 画图", "用 GPT 生成图片"), giving good coverage of terms users would actually say; not score 2 because common variations are already well represented.

Clear niche with explicit guardrails (use ONLY for named OpenAI/GPT, redirect generic requests to nanobanana-skill, exclude diagrams), making wrong-skill triggering unlikely; not score 2 because the scoping is explicit rather than merely somewhat specific.