paper-illustration-image2

Generate publication-quality academic illustrations through a local Codex app-server bridge that uses Codex native image generation. This is a separate experimental alternative to `paper-illustration`, intended for Claude Code users who want a GPT-image-style renderer without modifying the original skill.

Quality

61%

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

Fix and improve this skill with Tessl

tessl review fix ./skills/paper-illustration-image2/SKILL.md

Quality

Content

63%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.

The skill has an excellent, well-gated iterative workflow with explicit validation checkpoints and executable helper commands, but it is let down by verbose inline reference material and a broken/over-deep external reference, hurting token efficiency and progressive disclosure.

Suggestions

Move the full CVPR/ICLR/NeurIPS style guide and the multi-layer shell resolver into separate reference files (e.g. STYLE_GUIDE.md, RESOLVER.md) and link to them from SKILL.md, keeping only a concise summary inline.

Fix or remove the broken `../shared-references/integration-contract.md` reference so navigation is one level deep and all linked paths resolve.

Trim the basic visual-standard explanations Claude already knows (white background, sans-serif fonts, grayscale readability) to a tight bullet list, cutting padding without losing the arrow-direction guidance.

Dimension	Reasoning	Score
Conciseness	Mostly efficient with well-organized sections and a useful ASCII workflow diagram, but it re-explains basic academic-figure conventions (clean white background, sans-serif fonts, grayscale readability) that Claude already knows, and the multi-layer shell resolver plus bilingual comments add padding.	3 / 5
Actionability	Provides concrete, executable bash invocations of the helper (preflight, finalize, verify) with verified flags and a complete LaTeX snippet, with only minor gaps such as the MCP call parameters being described rather than shown as a ready template.	4 / 5
Workflow Clarity	The multi-stage workflow is explicitly sequenced (Step 0 through Step 7) with a preflight gate, a score-gated review/refine feedback loop (score >= 9, else loop back to Step 2), and a mandatory finalize+verify checkpoint, satisfying the validation/feedback-loop requirement for this iterative, effectively batch operation.	5 / 5
Progressive Disclosure	The body inlines substantial reference-grade material (the full CVPR/NeurIPS style guide and a 4-layer shell resolver) that belongs in separate files, the only external reference (`../shared-references/integration-contract.md`) does not resolve to an existing file, and the bundled helper is referenced by name without a clear one-level-deep navigation pointer.	2 / 5
	Total	14 / 20 Passed

Description

58%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 specific about the capability and clearly distinguishes the skill from its sibling, but it lacks an explicit 'Use when...' trigger clause and leans on internal product names rather than natural user phrasing, capping completeness and trigger-term quality.

Suggestions

Add an explicit 'Use when...' clause naming natural user triggers, e.g. 'Use when generating paper-ready architecture or method figures via a local Codex image bridge.'

Include natural synonyms and concrete figure types users would say (architecture diagrams, method illustrations, pipeline figures) instead of relying on internal names like 'Codex app-server'.

Lead with a concise action list ('Generate, refine, and review publication-quality academic illustrations') before the implementation detail about the bridge.

Dimension	Reasoning	Score
Specificity	Names the concrete capability (publication-quality academic illustration generation via a Codex app-server bridge using native image generation) and its positioning as an alternative to `paper-illustration`, but lacks a short list of distinct concrete actions like other 5-anchor examples.	4 / 5
Completeness	It clearly states what the skill does, but provides no explicit 'Use when...' trigger guidance, which the rubric caps at 3.	3 / 5
Trigger Term Quality	Uses natural terms ('publication-quality', 'academic illustrations', 'paper figures', 'GPT-image-style renderer') but omits common synonyms and file-extension triggers a user would naturally say, and relies on internal product names ('Codex app-server', 'paper-illustration') more than user-facing phrases.	3 / 5
Distinctiveness Conflict Risk	It is explicitly framed as a separate experimental alternative to `paper-illustration` targeting Claude Code users wanting a GPT-image-style renderer, giving it a clear niche with only minor overlap risk against the sibling skill.	4 / 5
	Total	14 / 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
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 1 suspicious	Warning

	Total	13 / 16 Passed

Repository: wanshuiyin/Auto-claude-code-research-in-sleep
Path: skills/paper-illustration-image2/SKILL.md
Commit: f4f20f9

Reviewed: 1 day 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.