borzoi

Predict genome-wide functional tracks (RNA-seq, CAGE, DNase, ChIP) from DNA sequence with Borzoi. Use this skill when: (1) Scoring the regulatory effect of a variant on expression/accessibility, (2) Generating predicted coverage tracks for a locus, (3) Prioritising non-coding variants by predicted track delta.

Quality

86%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

87%

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

The content is dense, executable, and well-structured with no filler — a strong example of token-efficient skill writing. The main gap is the absence of explicit validation/verification checkpoints in the remote-compute and variant-scoring workflows.

Suggestions

Add an explicit validation step after job completion, e.g. verify `tracks.npz` exists and check expected shape before calling `save_artifacts`, with a retry-on-failure branch.

For variant scoring, add a brief checkpoint confirming ref/alt windows share the same centring and length before comparing per-track deltas.

Clarify that the notification payload should be inspected for error/status fields before acting on `featured_files`.

Dimension	Reasoning	Score
Conciseness	The body is lean and assumes Claude's competence — it never explains what RNA-seq/DNase/ChIP are or how PyTorch works, and every section (prerequisites table, code, troubleshooting) earns its tokens, matching the 'lean and efficient' anchor.	3 / 3
Actionability	Provides copy-paste-ready executable code (`Borzoi.from_pretrained("johahi/borzoi-replicate-0")`), exact tensor shapes `(batch, 4, 524288) → (batch, tracks, 6144)`, and concrete fixes (`importlib.metadata.version("borzoi-pytorch")`, pad/crop to 524288 bp), satisfying 'fully executable code/commands; copy-paste ready'.	3 / 3
Workflow Clarity	The remote-compute flow is sequenced (submit_job → wait_for_notification → save_artifacts → attach_job) but lacks explicit validation checkpoints: it never instructs verifying job status or output existence before consuming `payload["featured_files"]`, and the ref/alt variant-scoring task has no verification step. Per the guideline, missing validation in batch/remote operations caps workflow clarity at 2; not 1 because the sequence is clearly present.	2 / 3
Progressive Disclosure	The body is under 50 lines with no bundle files and is well-organized into Prerequisites, How to run, Output format, Remote compute, and Troubleshooting sections; cross-skill pointers (remote-compute-ssh/modal) are one level deep and clearly signaled, fitting the simple-skills allowance for a 3.	3 / 3
	Total	11 / 12 Passed

Description

85%

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 strong: concrete capabilities, an explicit three-part 'Use when' trigger, and a clearly distinct niche. Its only weakness is trigger-term phrasing skewed toward specialist terminology rather than the everyday wording a broad user might say.

Suggestions

Add more colloquial trigger phrasings alongside the technical ones (e.g. 'predict RNA-seq / ChIP coverage from sequence', 'estimate variant impact on gene expression') so non-specialist users naturally match it.

Consider naming the model card artifact ('johahi/borzoi-replicate-0') is already covered in the body; keep the description focused on user-facing triggers rather than internal model IDs.

Dimension	Reasoning	Score
Specificity	Names the concrete capability ('Predict genome-wide functional tracks (RNA-seq, CAGE, DNase, ChIP) from DNA sequence') plus three distinct actions (variant effect scoring, coverage track generation, variant prioritisation), matching the 'lists multiple specific concrete actions' anchor.	3 / 3
Completeness	Explicitly states what it does (predict tracks from DNA) and when to use it via 'Use this skill when: (1)... (2)... (3)...', satisfying the 'clearly answers both what AND when with explicit triggers' anchor.	3 / 3
Trigger Term Quality	Terms like 'functional tracks', 'CAGE', 'DNase', 'regulatory effect', and 'non-coding variants' are accurate but specialized; it lacks the more common variations a general user might say, fitting 'some relevant keywords but missing common variations'. Not 3 because coverage of natural phrasings is incomplete; not 1 because the keywords are relevant rather than generic jargon.	2 / 3
Distinctiveness Conflict Risk	The Borzoi DNA→functional-track niche with variant-scoring triggers is a clear, narrow domain unlikely to conflict with other skills, matching the 'clear niche with distinct triggers' anchor.	3 / 3
	Total	11 / 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
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: UnicomAI/wanwu
Path: configs/microservice/bff-service/configs/agent-skills/claude-science/borzoi/SKILL.md
Commit: f618458

Reviewed: about 13 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.