stable-diffusion-image-generation

State-of-the-art text-to-image generation with Stable Diffusion models via HuggingFace Diffusers. Use when generating images from text prompts, performing image-to-image translation, inpainting, or building custom diffusion pipelines.

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

65%

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

A highly actionable skill rich in executable examples and reference tables, but it is verbose and monolithic: conceptual explanations and redundant workflow demos inflate the body, and multi-step/batch workflows lack validation checkpoints. Progressive disclosure is partially realized through two well-linked reference files, yet much detail remains inline.

Suggestions

Trim conceptual sections (architecture overview, pipeline-flow ASCII art) and the duplicate "Common workflows" demos, moving that detail into references/advanced-usage.md so SKILL.md stays a lean overview.

Add validation checkpoints to batch and risky workflows — e.g., assert output dimensions/dtype before saving and confirm GPU memory headroom after enabling optimizations — to introduce explicit validate-then-proceed feedback loops.

Replace the undefined get_canny_image(input_image) call in the ControlNet example with concrete cv2.Canny code so the example is fully executable as written.

Dimension	Reasoning	Score
Conciseness	The body is a ~510-line tutorial with conceptual sections Claude largely knows ("Three-pillar design", ASCII "Pipeline inference flow") and "Common workflows" that re-demonstrate techniques already shown; it is mostly actionable but could be tightened and offloaded to reference files, so it sits at 2 rather than the lean score-3 anchor.	2 / 3
Actionability	Provides numerous complete, copy-paste-ready code blocks plus concrete parameter and scheduler tables, matching the fully-executable score-3 anchor; the one undefined helper get_canny_image in the ControlNet example is a minor blemish that does not drop it to score-2's pseudocode level.	3 / 3
Workflow Clarity	Workflows list numbered steps but lack validation checkpoints or validate-fix-retry feedback loops, and batch generation is shown without any verification, which caps clarity at 2 per the batch/destructive-operations guideline.	2 / 3
Progressive Disclosure	Two real reference files (advanced-usage.md, troubleshooting.md) are clearly signaled one level deep, but the SKILL.md body itself holds a large amount of inline detail that could be split out, fitting the score-2 anchor where content that should be separate is inline.	2 / 3
	Total	9 / 12 Passed

Description

100%

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, third-person description that pairs a concrete capability statement with an explicit "Use when" trigger clause and natural user keywords. The only minor weakness is the "State-of-the-art" buzzword lead-in, which is mild over-claiming but does not undermine the concrete actions listed.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — "text-to-image generation", "image-to-image translation", "inpainting", and "building custom diffusion pipelines" — matching the score-3 anchor rather than the single-domain score-2 anchor.	3 / 3
Completeness	Explicitly states what it does ("State-of-the-art text-to-image generation with Stable Diffusion models via HuggingFace Diffusers") and when to use it via a clear "Use when..." clause, satisfying both what and when.	3 / 3
Trigger Term Quality	Covers natural terms a user would say ("text-to-image", "text prompts", "image-to-image translation", "inpainting") with good variation, beyond the sparse score-2 example.	3 / 3
Distinctiveness Conflict Risk	Scoped to a clear niche (Stable Diffusion / HuggingFace Diffusers) with distinct triggers unlikely to collide with other skills, rather than the generic score-2 "Works with document files" pattern.	3 / 3
	Total	12 / 12 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (521 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/llm-tools/stable-diffusion/SKILL.md
Commit: 52845c3

Reviewed: about 13 hours 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.