hugging-science

Use when the user is doing AI/ML work in a scientific domain such as biology, chemistry, physics, astronomy, climate, genomics, materials, medicine, ecology, energy, engineering, math, drug discovery, protein design, weather modeling, theorem proving, single-cell, or PDE solving. Hugging Science is a curated catalog of scientific datasets, models, blog posts, and interactive Spaces. This skill helps discover and use resources via `datasets`, `transformers`, the HF Inference API, `gradio_client`, and methodology citations.

Quality

87%

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

Quality

Content

85%

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

A well-structured, actionable skill body: clear five-step workflow, executable commands, and exemplary progressive disclosure with verified one-level-deep references. The main drag is conciseness — a promotional opening and some explanatory padding add tokens that don't all earn their place.

Suggestions

Trim the promotional framing in the opening paragraph ('much higher signal than generic search and the entries are pre-filtered for quality and openness') and other justificatory asides to reach the lean score-3 conciseness anchor.

Tighten the 'A few important things to remember' section by merging the two entries that overlap on catalog evolution/fallback so each bolded paragraph carries a distinct, non-redundant gotcha.

Dimension	Reasoning	Score
Conciseness	The body is mostly focused on non-obvious gotchas and avoids explaining basic concepts, but the promotional intro ('much higher signal than generic search', 'pre-filtered for quality and openness') and overall length (~130 lines with some explanatory padding) could be tightened to earn the score-3 'every token earns its place' anchor.	2 / 3
Actionability	It provides copy-paste-ready, executable guidance: concrete `fetch_catalog.py` commands with flags, exact fetch URLs, and a working `python-dotenv` snippet plus the `.env` format, while appropriately deferring detailed API code to focused reference files.	3 / 3
Workflow Clarity	The five-step Core workflow is clearly sequenced with numbered headers, and includes explicit checkpoints — 'present the top 2–3 candidates to the user... proceed once they choose' and 'If a URL 404s, refetch the topic file' — which satisfy the validation/feedback-loop anchor for this non-destructive discovery workflow.	3 / 3
Progressive Disclosure	SKILL.md is a well-organized overview pointing to six real, one-level-deep bundle files (5 references + 1 script, all verified present), each signaled inline at the relevant step and re-listed in a 'Bundled resources' section for easy navigation.	3 / 3
	Total	11 / 12 Passed

Description

90%

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 description: explicit what/when triggers with a rich, natural set of scientific-domain keywords and a clearly delineated niche. The only weakness is that the action verbs are somewhat generic ('discover and use') even though the underlying tools are concretely named.

Suggestions

Replace the generic verbs 'discover and use' with a few concrete actions (e.g., 'load scientific datasets, run domain models locally or via the Inference API, call interactive Spaces, cite methodology blogs') to lift the specificity anchor to 3.

Consider pairing each named tool with the concrete action it performs (e.g., 'load datasets via `datasets`, run models via `transformers`') so the actions read as specific rather than abstract.

Dimension	Reasoning	Score
Specificity	It names many domains and concrete tools (`datasets`, `transformers`, the HF Inference API, `gradio_client`) but the actual verbs are generic — 'helps discover and use resources' and 'methodology citations' — rather than multiple distinct concrete actions like the score-3 anchor (extract/fill/merge).	2 / 3
Completeness	It explicitly answers both what ('Hugging Science is a curated catalog of scientific datasets, models, blog posts, and interactive Spaces... helps discover and use resources') and when ('Use when the user is doing AI/ML work in a scientific domain such as...'), with an explicit trigger clause.	3 / 3
Trigger Term Quality	The 'Use when...' clause enumerates a broad set of natural terms a user would actually say — biology, chemistry, physics, genomics, drug discovery, protein design, weather modeling, theorem proving, single-cell, PDE solving — giving excellent coverage of scientific-domain triggers.	3 / 3
Distinctiveness Conflict Risk	The scientific-domain scoping is a clear, narrow niche with distinct triggers unlikely to fire for generic ML skills; it even contrasts itself against 'generic ML' tasks to reduce overlap.	3 / 3
	Total	11 / 12 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: K-Dense-AI/scientific-agent-skills
Path: skills/hugging-science/SKILL.md
Commit: 757b63b

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.