ml-paper-writing

Write publication-ready ML/AI papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM. Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions. For systems venues (OSDI, NSDI, ASPLOS, SOSP), use systems-paper-writing instead.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Quality

Content

77%

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

The content is highly actionable and well-sequenced with strong validation checkpoints, but it is too long and repetitive for its token budget and suffers from broken bundle references (templates/, systems-paper-writing) plus inline material that belongs in the existing reference files. Conciseness and progressive disclosure are the weakest dimensions.

Suggestions

Consolidate the three separate 'never hallucinate citations' sections into one and reduce the ~8 systems-paper-writing cross-references to a single pointer, cutting hundreds of redundant lines.

Move the inline conference page-limit/requirement tables and template quick-reference tables into the existing references/ files (e.g. checklists.md, a new templates.md) so the SKILL.md body is a lean overview.

Fix or remove broken bundle references: the body repeatedly links to templates/ and templates/README.md and to ../systems-paper-writing/, none of which exist in the skill.

Dimension	Reasoning	Score
Conciseness	The body runs ~980 lines and repeats itself: three separate citation-hallucination warning sections and roughly eight cross-references to systems-paper-writing restate the same points. It is mostly actionable rather than explanatory, so it is above level 1, but the redundancy and length keep it below the 'every token earns its place' level 3.	2 / 3
Actionability	It provides copy-paste-ready, executable guidance — a real `doi_to_bibtex` Python function, Semantic Scholar search code, `latexmk -pdf main.tex` and `cp -r templates/` shell commands, and concrete LaTeX/table snippets — matching the fully-executable level-3 anchor.	3 / 3
Workflow Clarity	Workflows are explicitly sequenced with validation checkpoints and feedback loops: 'Verify template compiles as-is (before any changes)', 'Verify paper exists in 2+ sources', and 'If ANY step fails → mark as placeholder, inform scientist', with checklists for each multi-step process — matching the level-3 anchor.	3 / 3
Progressive Disclosure	The five references/ files are well-signaled and one level deep, but the body also links to a templates/ directory and ../systems-paper-writing/ that do not exist in the bundle, and keeps large reference-style tables (conference page limits, template quick-reference) inline in an already-massive overview. This fits 'some structure but could be better organized; content that should be separate is inline' better than the clean level-3 split.	2 / 3
	Total	10 / 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.

The description is third-person, specific, and complete: it states what the skill does, gives an explicit 'Use when' trigger clause with natural venue keywords, and cleanly distinguishes itself from the systems-paper-writing skill. All four dimensions land at the top of the scale.

Dimension	Reasoning	Score
Specificity	The description lists multiple concrete actions — 'drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions' — matching the 'lists multiple specific concrete actions' anchor, not the single-action level 2.	3 / 3
Completeness	It explicitly answers both what ('Write publication-ready ML/AI papers for...') and when via an explicit 'Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions' clause, satisfying the level-3 anchor with explicit triggers.	3 / 3
Trigger Term Quality	It covers natural terms the target audience actually says: venue names (NeurIPS, ICML, ICLR, ACL, AAAI, COLM), 'drafting papers from research repos', 'verifying citations', and 'camera-ready submissions' — good coverage rather than the partial set at level 2.	3 / 3
Distinctiveness Conflict Risk	It carves a clear niche (ML/AI conferences) and actively routes the overlapping systems case elsewhere ('For systems venues (OSDI, NSDI, ASPLOS, SOSP), use systems-paper-writing instead'), making wrong-skill triggering unlikely.	3 / 3
	Total	12 / 12 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
skill_md_line_count	SKILL.md is long (983 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 2 missing, 7 suspicious	Warning

	Total	13 / 16 Passed

Repository: Orchestra-Research/AI-Research-SKILLs
Path: 20-ml-paper-writing/ml-paper-writing/SKILL.md
Commit: 773a529

Reviewed: about 11 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.