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. Includes LaTeX templates, reviewer guidelines, and citation verification workflows.

Quality

71%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./backend/cli/skills/writing/ml-paper-writing/SKILL.md

Quality

Content

56%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 content is information-rich and largely actionable with clear numbered workflows and validation checkpoints, but it suffers from significant verbosity and duplication (especially around citation hallucination warnings and the source bibliography) and references a `templates/` directory that is not present in the bundle. Tightening the repeats and moving detail-heavy sections into the existing references would materially improve token efficiency.

Suggestions

Consolidate the three repeated 'never hallucinate citations / ~40% error rate' warnings into a single CRITICAL section and have the later Citation Workflow reference it rather than restating it.

Remove the duplicate external-source bibliography: keep one (either the 'Sources Behind This Guidance' table or the 'Key External Sources' list) and point the other to references/sources.md.

Either include the referenced `templates/` directory in the bundle or remove the five `templates/` pointers (including `templates/README.md`) so every cited path resolves to a real file.

Dimension	Reasoning	Score
Conciseness	At ~960 lines the body is noticeably verbose and repeats itself: the 'never hallucinate / ~40% error rate / never write BibTeX from memory' warning appears three separate times (sections 'CRITICAL: Never Hallucinate Citations', 'Citation Workflow', and the closing summary), the external-source bibliography is listed twice (the sources table and 'Key External Sources'), and much of the checklist/template prose restates principles Claude already knows.	2 / 5
Actionability	Mostly executable guidance: concrete bash/latex/python snippets (e.g. `cp -r templates/neurips2025/`, the `doi_to_bibtex` Python function, `latexmk -pdf main.tex`) cover common cases with only minor gaps such as the un-verified `templates/` directory and the Semantic Scholar snippet lacking the install step.	4 / 5
Workflow Clarity	Numbered workflows with checkboxes and feedback loops are present (draft -> feedback -> revise, validate -> fix -> retry for citations), with explicit 'MANDATORY' validation steps; the main gap is that the template-compile workflow lacks an explicit 'compile failed -> diagnose -> recompile' loop and dangling template references weaken the verified-checkpoint chain.	4 / 5
Progressive Disclosure	Structure is reasonable and references are one level deep with clear signaling (a reference table maps each references/*.md to contents, and bundle files exist), but the SKILL.md itself still inlines large blocks that belong in references (full citation workflow, conference requirement tables, template pitfalls) and it repeatedly points to a `templates/` directory that does not exist in the bundle.	3 / 5
	Total	13 / 20 Passed

Description

87%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.

A strong, well-constructed description: it states concrete capabilities, names the target venues, and provides an explicit 'Use when' trigger clause with natural keywords. Minor expansion to include synonyms like 'manuscript' or 'rebuttal' would round out trigger coverage.

Dimension	Reasoning	Score
Specificity	Names the domain and several concrete actions ('drafting papers from research repos', 'structuring arguments', 'verifying citations', 'preparing camera-ready submissions') plus bundle contents, with only minor coverage gaps (no mention of figure/table generation or rebuttal writing).	4 / 5
Completeness	Clearly answers both 'what' ('Write publication-ready ML/AI papers for [venues]') and 'when' with an explicit 'Use when...' clause listing concrete trigger phrases, matching the top anchor.	5 / 5
Trigger Term Quality	Good natural keyword coverage ('drafting papers', 'research repos', 'verifying citations', 'camera-ready submissions', conference names) that users would actually say, though synonyms like 'manuscript', 'rebuttal', or 'revision' are missing.	4 / 5
Distinctiveness Conflict Risk	Occupies a clear niche (top-venue ML/AI paper writing with named conferences and citation-verification workflows) with distinct, specific triggers that minimize conflict risk with general writing skills.	5 / 5
	Total	18 / 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
skill_md_line_count	SKILL.md is long (960 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	Warning

	Total	13 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/writing/ml-paper-writing/SKILL.md
Commit: 1d182e9

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.