CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

non-tumor-mechanism-guided-diagnostic-ml-research-planner

Generates complete conventional non-oncology diagnostic machine-learning research designs from a user-provided disease context, optional mechanism theme, and validation direction. Use when a study centers on disease-vs-control transcriptome comparison, optional mechanism-gene restriction, feature shrinkage, diagnostic model construction, ROC / calibration / DCA evaluation, interpretation layers, and orthogonal validation. Covers five study patterns and always outputs Lite / Standard / Advanced / Publication+ with a recommended primary plan, stepwise workflow, figure plan, validation hierarchy, minimal executable version, publication upgrade path, and strictly verified literature retrieval.

Quality

87%

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

81%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 highly actionable, well-sequenced planning skill: exact output templates, explicit validation checkpoints, and clean one-level reference links. The main costs are token efficiency — heavy duplication between the step sections and the Hard Rules list — and three orphaned bundle files with no navigation path from the main file.

Suggestions

Consolidate mandates stated twice: fold Hard Rules 11–13, 15, 17, 20–23 into the step/section text they duplicate (or vice versa) so each rule appears once.

Link the three unreferenced bundle files (references/analysis-modules.md, references/method-library.md, references/workflow-step-template.md) from the relevant steps so they are discoverable.

Trim the repeated endpoint-formula examples in Step 5 and Section C.5 to a single canonical list.

DimensionReasoningScore

Conciseness

The body is dense, instruction-only content with no conceptual over-explanation, but the 23 Hard Rules substantially duplicate the step sections: reference-verification rules appear in Step 4.5, Hard Rules 11–13, and Section I; Section G subset rules in Section G and Hard Rules 15/17; the Dataset Disclaimer in Section D and Hard Rule 22. This is mostly efficient but could be tightened by consolidating each mandate in one place.

3 / 5

Actionability

Guidance is fully executable for an instruction-only skill: exact output sections A–J with per-section required elements, a verbatim out-of-scope redirect script, a verbatim Dataset Disclaimer line, minimum retrieval counts (2–4 background, 2–4 method, 1–2 precedent), and the DESeq2-vs-limma selection rule keyed to input data type — Claude knows exactly what to produce.

5 / 5

Workflow Clarity

Seven explicitly ordered steps with a mandatory pre-output validation checkpoint (Step 5 dependency consistency: 'If dependency fails, remove or downgrade the downstream claim rather than silently keeping it'), an out-of-scope redirect, a fallback plan in Section J, and a downgrade map — feedback loops cover the risky steps.

5 / 5

Progressive Disclosure

The body links 5 reference files, all of which exist in references/ and are one level deep with no nested references, clearly signaled ('→ Detailed pattern logic: [references/study-patterns.md](references/study-patterns.md)'). However, 3 of the 8 bundle files (analysis-modules.md, method-library.md, workflow-step-template.md) are never referenced from SKILL.md, leaving bundle content undiscoverable — a minor organization gap.

4 / 5

Total

17

/

20

Passed

Description

88%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: third-person, concrete about both outputs and trigger conditions, and firmly niche-scoped. Its only weaknesses are missing natural synonyms (biomarker, GEO, DEG) and no explicit differentiation from sibling research-planner skills.

Suggestions

Add high-frequency natural trigger terms users would actually type — 'biomarker', 'GEO dataset', 'DEG', 'diagnostic signature' — to the 'Use when' clause.

Add one boundary phrase distinguishing this skill from closely related sibling planners (e.g., 'use for diagnostic-model papers, not hub-gene or prognostic-survival studies').

DimensionReasoningScore

Specificity

The description lists multiple concrete actions and deliverables ('Generates complete... research designs', 'always outputs Lite / Standard / Advanced / Publication+ with a recommended primary plan, stepwise workflow, figure plan, validation hierarchy, minimal executable version, publication upgrade path, and strictly verified literature retrieval'), giving comprehensive coverage of the skill's outputs with no vague padding.

5 / 5

Completeness

It explicitly answers what ('Generates complete conventional non-oncology diagnostic machine-learning research designs from a user-provided disease context, optional mechanism theme, and validation direction') and when ('Use when a study centers on disease-vs-control transcriptome comparison, ... ROC / calibration / DCA evaluation ...'), with concrete trigger phrases in third-person voice.

5 / 5

Trigger Term Quality

Domain-natural terms like 'diagnostic model', 'ROC / calibration / DCA', 'disease-vs-control transcriptome comparison', and 'feature shrinkage' are present, but common variations a user would actually say — 'biomarker', 'GEO', 'DEG', 'machine learning' spelled out — are missing, so coverage is good rather than comprehensive.

4 / 5

Distinctiveness Conflict Risk

'Conventional non-oncology diagnostic machine-learning research designs' carves out a clear niche, but the same bundle family contains closely related planner skills (e.g., the hub-gene planner referenced in the body) and the description does not draw an explicit boundary against them, leaving minor overlap risk.

4 / 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

awesome-med-research-skills/Protocol Design/non-tumor-mechanism-guided-diagnostic-ml-research-planner/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.