paper-write
Draft LaTeX paper section by section from an outline. Use when user says \"写论文\", \"write paper\", \"draft LaTeX\", \"开始写\", or wants to generate LaTeX content from a paper plan.
64
76%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Critical
Do not install without reviewing
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/skills-codex/paper-write/SKILL.mdQuality
Discovery
89%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a solid skill description that clearly communicates its purpose and provides explicit trigger guidance in both English and Chinese. Its main weakness is that the 'what' portion could be more specific about the concrete actions performed (e.g., generating sections, handling citations, formatting equations). Overall, it performs well on completeness and distinctiveness.
Suggestions
Add more specific concrete actions to the 'what' portion, e.g., 'generates title, abstract, introduction, methodology, results, and conclusion sections with proper LaTeX formatting and citation placeholders'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names the domain (LaTeX paper drafting) and a key action (draft section by section from an outline), but doesn't list multiple specific concrete actions like formatting, adding citations, generating bibliographies, or creating figures. | 2 / 3 |
Completeness | Clearly answers both 'what' (draft LaTeX paper section by section from an outline) and 'when' (explicit 'Use when' clause with specific trigger phrases and a general condition about generating LaTeX from a plan). | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural trigger terms in both English and Chinese: '写论文', 'write paper', 'draft LaTeX', '开始写', plus the conceptual trigger 'generate LaTeX content from a paper plan'. These are terms users would naturally say. | 3 / 3 |
Distinctiveness Conflict Risk | The combination of LaTeX, paper drafting, section-by-section from an outline, and bilingual trigger terms creates a clear niche that is unlikely to conflict with general writing skills or generic LaTeX formatting skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
62%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a highly actionable and well-structured skill with an excellent multi-step workflow including validation checkpoints, fallback chains, and comprehensive final checks. However, it is severely over-long and repetitive — key rules like citation style differences are restated 3-4 times, writing quality guidance is duplicated between Step 5 and the Writing Quality Reference section, and many concepts Claude already knows are explained unnecessarily. The content would benefit greatly from aggressive trimming and splitting detailed reference material into separate files.
Suggestions
Reduce redundancy by stating venue-specific rules (IEEE vs natbib, page limit differences) in exactly one place and referencing that location elsewhere instead of repeating them in Constants, Templates, Step 3, and Key Rules.
Move the detailed section-specific writing guidelines (§0-§5 and Appendix) and the 5-pass quality audit into a separate reference file (e.g., section-guidelines.md and quality-pass.md) and reference them from the main workflow.
Remove the Acknowledgements section and explanations of basic concepts (e.g., what an abstract should contain, how paragraphs work) — these consume tokens without adding value for Claude.
Consolidate the 'Writing Quality Reference' section with Step 5 — currently the de-AI patterns and writing principles appear in both places with overlapping content.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~400+ lines. It explains concepts Claude already knows (what abstracts are, how paragraphs work, basic LaTeX structure), includes lengthy acknowledgements, repeats rules multiple times (e.g., IEEE vs natbib citation style is stated in at least 4 places), and contains extensive boilerplate that could be dramatically condensed. The writing quality pass section alone restates principles that are already in the referenced shared files. | 1 / 3 |
Actionability | The skill provides highly concrete, executable guidance throughout: specific curl commands for DBLP/CrossRef fetching, complete LaTeX template snippets for each venue, exact file structures, Python code patterns for bib cleaning, specific spawn_agent syntax for cross-review, and detailed checklists. Nearly everything is copy-paste ready. | 3 / 3 |
Workflow Clarity | The 8-step workflow is clearly sequenced (Step 0 through Step 8) with explicit validation checkpoints: backup before overwrite (Step 0), theory consistency pass (Step 3.5), 5-pass quality audit (Step 5), cross-review with reviewer model (Step 6), reverse outline test (Step 7), and a comprehensive final checklist (Step 8). Error recovery is addressed (e.g., DBLP fallback chain, file size retry with bash). | 3 / 3 |
Progressive Disclosure | The skill references external files (writing-principles.md, venue-checklists.md, citation-discipline.md, templates/) appropriately, but the main SKILL.md itself is monolithic — it inlines enormous amounts of detail that could be split into separate reference files (e.g., venue templates, section-specific guidelines, the full quality pass details). No bundle files were provided, so the referenced paths cannot be verified. | 2 / 3 |
Total | 9 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Commit
a425a71
- Reviewed
Table of Contents
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.