CtrlK
BlogDocsLog inGet started
Tessl Logo

deepspot-m

Generate transcriptome-wide virtual spatial transcriptomics from H&E histology with DeepSpot-M. Use when you need spatial gene expression in log1p-CPM for 224x224 tiles at about 20x, want to query protein-coding genes by symbol instead of a fixed panel, or want to run prediction across a whole slide after tiling with histolab.

75

Quality

93%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

86%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 body that leads with a tight overview and supplies executable code for install, model access, quick start, tile validation, and optional-dependency loading, while correctly offloading detailed API and whole-slide mechanics to two real one-level-deep references. The only meaningful gap is the absence of an inline validation/retry loop for the batch whole-slide workflow.

Suggestions

Add an explicit validate→fix→retry checkpoint into the whole-slide workflow steps in SKILL.md (e.g. verify batch shapes/coordinates before concatenation), rather than leaving it only in the reference, since batch operations should have an inline feedback loop.

Tighten the Overview architecture paragraph (the TCGA '28,664 slides across 32 cancer types' figure and the enumerated frozen-embedding list) to the minimum needed for the gene-router/source concept, moving the rest to references/api.md.

In the Embedding sources table, fold the one-line 'Each gives a different view of gene identity' guidance into the table or a single sentence to remove a lightly padded explanation line.

DimensionReasoningScore

Conciseness

Lean and mostly efficient; the architecture paragraph (LoRA backbone, cross-attention decoder, router hypernetwork, the five frozen embeddings) is genuinely model-specific rather than Claude-known padding, though a few lines (e.g. the TCGA atlas figure) could be trimmed without loss. Not a 5 because minor trimming remains possible.

4 / 5

Actionability

Fully executable, copy-paste-ready guidance throughout: `uv pip install deepspotm==1.0.0`, `huggingface-cli login`, the quick-start `from_pretrained`/`predict_genes` snippet, a `require_tile` validator, and a `load_deepspotm` optional-dependency loader covering the common cases.

5 / 5

Workflow Clarity

The whole-slide workflow gives a clear 4-step sequence and the tile-requirements section includes an explicit boundary validation checkpoint (`require_tile` raises on wrong size). Not a 5 because the batch workflow's validate→fix→retry loop is deferred to the reference rather than stated inline.

4 / 5

Progressive Disclosure

Clear overview with well-signaled one-level-deep references to real files: `references/api.md` (full call surface) and `references/whole_slide.md` (worked loop), both verified present, with detailed content appropriately split out of the body.

5 / 5

Total

18

/

20

Passed

Description

100%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 precise, well-constructed description that states concrete capabilities, uses third-person imperative voice, and pairs a clear 'what' with an explicit multi-condition 'Use when' trigger. It is distinctive within a narrow niche and free of vague fluff or over-claims.

DimensionReasoningScore

Specificity

Lists multiple specific concrete actions — 'Generate transcriptome-wide virtual spatial transcriptomics from H&E histology', 'query protein-coding genes by symbol instead of a fixed panel', and 'run prediction across a whole slide after tiling with histolab' — giving comprehensive coverage of the skill's capabilities.

5 / 5

Completeness

Clearly answers 'what' ('Generate transcriptome-wide virtual spatial transcriptomics from H&E histology with DeepSpot-M') and 'when' via an explicit 'Use when...' clause with three concrete triggering conditions.

5 / 5

Trigger Term Quality

Comprehensive coverage of natural and technical terms users would say, including synonyms: 'spatial gene expression', 'log1p-CPM', '224x224 tiles at about 20x', 'protein-coding genes by symbol', 'whole slide', and 'tiling with histolab'.

5 / 5

Distinctiveness Conflict Risk

Occupies a clear, narrow niche (virtual spatial transcriptomics from H&E via DeepSpot-M) with distinct triggers that are unlikely to fire for any other skill, giving minimal conflict risk.

5 / 5

Total

20

/

20

Passed

Validation

100%

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

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
K-Dense-AI/scientific-agent-skills
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.