CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

geniml

Machine learning toolkit for genomic interval (BED) data; use it when you need to tokenize BED collections and train embeddings for regions/cells/labels, build consensus peak universes, or run similarity search and downstream ML on chromatin accessibility datasets.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

68%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.

Well-structured, actionable overview with real one-level-deep reference files and two complete example pipelines. Main weaknesses are the absence of validation checkpoints in the batch tokenization/training workflows and the vague BEDspace operational guidance.

Suggestions

Add explicit validation checkpoints between pipeline stages (e.g., verify the tokens folder is non-empty and spot-check universe peak count after 'geniml universe build') to lift workflow clarity above the batch-operation cap.

Provide a concrete BEDspace example (install command for StarSpace plus a representative training invocation) instead of the vague 'its path provided/configured' instruction.

Replace the hedged 'Additional details are commonly documented in' blockquote with inline per-feature links (e.g., '**Region2Vec**: See [references/region2vec.md](references/region2vec.md)') so navigation to each reference is clearly signaled.

DimensionReasoningScore

Conciseness

The body is efficient — lean code blocks, terse dependency lists, and practical parameter notes like 'lower is stricter; overly strict thresholds can reduce coverage' — but a few prose passages ('Core idea' explanations, hedged wording like 'Additional details are commonly documented in') could be trimmed, fitting anchor 4 rather than the every-token-earns-its-place lean of anchor 5.

4 / 5

Actionability

Two complete executable pipelines (bash universe build, Python tokenize→train→evaluate; and the Scanpy scEmbed workflow) plus concrete install commands make the guidance mostly copy-paste ready, but the BEDspace workflow offers no example command ('its path provided/configured' stays vague), a minor gap consistent with anchor 4.

4 / 5

Workflow Clarity

Sequences are clearly ordered (coverage → universe → tokenize → train → evaluate, with numbered in-code comments), but there are no intermediate validation checkpoints for batch operations like tokenizing entire folders of BED files — the rubric's batch-operation cap holds workflow clarity at 3, and this matches anchor 3 ('steps listed but validation gaps; checkpoints missing or implicit') better than anchor 4.

3 / 5

Progressive Disclosure

All five referenced files (references/region2vec.md, bedspace.md, scembed.md, consensus_peaks.md, utilities.md) exist, are one level deep, and cover the major features; structure is good with clear sections, though the reference pointer is a single hedged blockquote rather than inline per-feature links and some per-method detail that the references likely cover is inlined — minor organization gaps fitting anchor 4.

4 / 5

Total

15

/

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 description that explicitly states both capabilities and use-when triggers with domain-specific vocabulary. Its only deductions are the second-person 'you need to' phrasing and a few missing natural synonyms (ATAC-seq, ChIP-seq).

DimensionReasoningScore

Specificity

The description lists multiple concrete actions — 'tokenize BED collections and train embeddings for regions/cells/labels, build consensus peak universes, or run similarity search and downstream ML' — matching the anchor-5 example's breadth, but the second-person phrasing 'use it when you need to' incurs the rubric's −1 voice penalty, landing it at 4 rather than 5.

4 / 5

Completeness

It clearly answers 'what' ('Machine learning toolkit for genomic interval (BED) data' plus the enumerated capabilities) and 'when' with the explicit trigger clause 'use it when you need to...', directly matching the anchor-5 example pattern.

5 / 5

Trigger Term Quality

Good keyword coverage including 'genomic interval (BED) data', 'consensus peak universes', 'similarity search', and 'chromatin accessibility', but common natural phrasings users would say such as 'ATAC-seq', 'ChIP-seq', or 'peak set' are missing, which fits anchor 4 rather than the comprehensive synonym coverage of anchor 5.

4 / 5

Distinctiveness Conflict Risk

The genomic-interval/BED ML niche is highly specific with distinct trigger terms, so conflict risk with other skills is minimal — a clear anchor-5 fit; it is well above anchor 4's 'minor overlap risk with closely related skills'.

5 / 5

Total

18

/

20

Passed

Validation

93%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository

aipoch/medical-research-skills

Path

scientific-skills/Data Analysis/geniml/SKILL.md

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.