geniml

This skill should be used when working with genomic interval data (BED files) for machine learning tasks. Use for training region embeddings (Region2Vec, BEDspace), single-cell ATAC-seq analysis (scEmbed), building consensus peaks (universes), or any ML-based analysis of genomic regions. Applies to BED file collections, scATAC-seq data, chromatin accessibility datasets, and region-based genomic feature learning.

Quality

81%

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

71%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.

A well-structured, actionable skill body with good progressive disclosure into verified reference files. Its main weaknesses are redundant capability coverage across three sections and a small executable typo in the install command.

Suggestions

Collapse the redundancy between 'Core Capabilities' workflows, 'Common Workflows' code, and 'When to Use Which Tool' into a single per-capability block to recover token budget.

Fix the install commands: 'uv uv pip install geniml' should be 'uv pip install geniml' (and likewise for the [ml] and git variants) so they run as written.

Add an explicit validation feedback loop to the Universe Building workflow (e.g., after `geniml universe evaluate`, state what to check and how to remediate before relying on the universe for tokenization).

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient code and short descriptions, but the five capabilities are covered three times over (Core Capabilities workflows, Common Workflows code, and the "When to Use Which Tool" section), which is unnecessary repetition that could be tightened.	3 / 5
Actionability	It provides concrete, mostly copy-paste-ready Python pipelines and CLI commands with real parameters, but the installation block reads "uv uv pip install geniml" — a doubled token that fails if copied verbatim — a minor executable gap.	4 / 5
Workflow Clarity	Multi-step pipelines (Basic Region Embedding, scATAC, Universe Building) are clearly numbered and include evaluate steps as checkpoints, but there are no explicit validate→fix→retry feedback loops for the batch universe-building operation.	4 / 5
Progressive Disclosure	The SKILL.md is a clear overview that signals five real, one-level-deep reference files (region2vec.md, bedspace.md, scembed.md, consensus_peaks.md, utilities.md, all present in ./references/), with content appropriately split and easy to navigate.	5 / 5
	Total	16 / 20 Passed

Description

92%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, specific description that clearly conveys both capability and trigger conditions for a well-defined genomics ML niche. The only minor gap is the absence of explicit file extensions among the trigger terms.

Dimension	Reasoning	Score
Specificity	The description lists multiple concrete, named capabilities — "training region embeddings (Region2Vec, BEDspace), single-cell ATAC-seq analysis (scEmbed), building consensus peaks (universes), or any ML-based analysis of genomic regions" — giving comprehensive coverage of the skill's actions rather than vague language.	5 / 5
Completeness	It explicitly answers both what (train embeddings, scATAC analysis, consensus peaks, ML analysis) and when ("This skill should be used when working with genomic interval data (BED files) for machine learning tasks. Use for... Applies to...") with concrete trigger phrases.	5 / 5
Trigger Term Quality	It includes strong natural terms users would say ("BED files", "single-cell ATAC-seq"/"scATAC-seq", "chromatin accessibility", "consensus peaks") with synonyms, but stops short of explicit file extensions like ".bed", so a few natural terms are missing.	4 / 5
Distinctiveness Conflict Risk	The genomic-interval ML niche is highly specific with distinct triggers (BED files, scATAC-seq, chromatin accessibility, region embeddings), giving minimal overlap risk with other skills.	5 / 5
	Total	19 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/ml-training/geniml/SKILL.md
Commit: 543b17f

Reviewed: 5 days 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.