variant-pathogenicity-predictor

Integrate REVEL, CADD, PolyPhen scores to predict variant pathogenicity.

Quality

62%

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

Fix and improve this skill with Tessl

tessl review fix ./scientific-skills/Data Analysis/variant-pathogenicity-predictor/SKILL.md

Quality

Content

42%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.

The body is dominated by generic packaging boilerplate at the expense of domain instruction, and its most important actionable content is wrong: the documented CLI (--variant/--vcf/--gene/--scores/--output), usage examples, and promised outputs (report.json, ACMG interpretation, confidence assessment) do not match the actual script, which only accepts --revel/--cadd/--polyphen/--demo and prints a composite score to stdout. The workflow and structure are present but templated, with broken internal cross-references.

Suggestions

Rewrite the Usage and Parameters sections to match the actual script interface (--revel, --cadd, --polyphen, --demo) or extend scripts/main.py to support the documented --variant/--vcf/--gene/--scores flags, and remove claims of outputs the script does not produce (ACMG interpretation, confidence assessment, report.json).

Cut the generic template sections (Security Checklist, Risk Assessment, Lifecycle Status, Evaluation Criteria, Response Template, Output Requirements, Implementation Details) that contain no domain-specific content, reducing the file to the workflow, real CLI, and score-integration details.

Remove or fix the broken "See `## X` above" cross-references in Dependencies, Example Usage, and Implementation Details, which point to sections that appear later in the file.

Dimension	Reasoning	Score
Conciseness	The ~190-line body is heavily padded with generic template sections ("Security Checklist", "Risk Assessment", "Lifecycle Status", "Evaluation Criteria", "Response Template", "Output Requirements") that add no domain knowledge, plus filler like "See `## Prerequisites` above for related details". This matches "Noticeably verbose; several unnecessary explanations or padded sections"; not 1 because it never explains concepts Claude already knows and does contain real executable commands.	2 / 5
Actionability	Some concrete guidance works ("python -m py_compile scripts/main.py", "python scripts/main.py --help", the run plan), but the primary Usage examples ("--variant \"chr17:43094692:G:A\" --gene \"BRCA1\"", "--vcf variants.vcf --output report.json") and the Parameters section document an interface that does not exist in scripts/main.py (actual flags are --revel, --cadd, --polyphen, --demo), so the key examples are not executable. Fits "Some concrete guidance but incomplete... missing key details"; not 2 because several genuinely executable verification commands are present.	3 / 5
Workflow Clarity	The Workflow section lists a sequenced 5-step process with an early scope-validation checkpoint and a fallback path, but the steps are generic template language, and validation of actual execution results is only implicit ("Review the generated output"). The documented VCF batch processing has no output verification step, so the batch-operation cap of 3 applies. Not 4 because checkpoints concern request scope rather than domain execution validation.	3 / 5
Progressive Disclosure	The bundle is appropriately shallow (SKILL.md plus a real, clearly-signaled "scripts/main.py" reference), but the internal navigation is broken: "See `## Prerequisites` above", "See `## Usage` above", and "See `## Workflow` above" appear in sections that precede the referenced sections, and substantial generic template content is inlined rather than split out. Fits "Some structure but could be better organized; references present but not clearly signaled"; not 4 because the misplaced cross-references actively confuse navigation.	3 / 5
	Total	11 / 20 Passed

Description

65%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 terse, domain-specific description that clearly states what the skill does, with excellent distinctiveness from tool-name triggers. Its main weakness is the complete absence of a "Use when..." clause, which caps completeness and limits natural-language discoverability for users who don't already know the tool names.

Suggestions

Append an explicit trigger clause, e.g. "Use when a user asks whether a variant is pathogenic or benign, or mentions REVEL, CADD, PolyPhen, VCF files, or ACMG classification."

Add one or two natural synonyms such as "classify" or "interpret variant" to broaden trigger-term coverage beyond users who already name the three tools.

Dimension	Reasoning	Score
Specificity	"Integrate REVEL, CADD, PolyPhen scores to predict variant pathogenicity" names the domain and specific tools with essentially one action pipeline (integrate scores -> predict), matching "Names domain and 1-2 concrete actions, but not comprehensive". Not 4 because it does not list several distinct actions (no classification, VCF processing, or reporting verbs).	3 / 5
Completeness	The "what" is clear (integrate three scoring tools to predict pathogenicity) but there is no "Use when..." clause or equivalent trigger guidance, which caps completeness at 3 per the judging guidelines. Not 4 because the "when" is entirely absent rather than merely implicit or underspecified.	3 / 5
Trigger Term Quality	"REVEL, CADD, PolyPhen" and "predict variant pathogenicity" are exactly the natural terms a clinical-genetics user would say, giving good keyword coverage; only a few natural variations are missing ("classify variant", "interpret variant", "VCF", "pathogenic vs benign"). Not 5 because there is no synonym or file-extension coverage beyond the three tool names.	4 / 5
Distinctiveness Conflict Risk	The combination of REVEL, CADD, and PolyPhen for variant pathogenicity is a clear niche with distinct trigger terms and minimal risk of firing for an unrelated skill. Not 4 because the highly specific tool names leave essentially no overlap risk even against closely related genomics skills.	5 / 5
	Total	15 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
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/variant-pathogenicity-predictor/SKILL.md
Commit: 686e09d

Reviewed: 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.