Name: tooluniverse-protein-sae-variant-interpretation
Rating: 70.39999999999999 (1 reviews)
Author: mims-harvard

Blog Docs Log in Get started

tooluniverse-protein-sae-variant-interpretation

Interpret a missense variant via ESMC-6B Sparse Autoencoder (SAE) feature activations. For a given protein + variant, computes which interpretable SAE features (catalytic, ligand-binding, PTM, structural motif, domain, etc.) are lost or gained at the mutation site. Use when standard pathogenicity scores (AlphaMissense, ClinVar) say a variant is damaging but you need a MECHANISTIC explanation — e.g. 'why is this variant LoF?' Complements (does not replace) variant-interpretation and variant-to-mechanism skills, which focus on ACMG classification or regulatory mechanism.

Quality

85%

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

The canonical home for this skill is tooluniverse-protein-sae-variant-interpretation in mims-harvard/ToolUniverse

Quality

Content

82%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, highly actionable skill body with executable code, explicit input validation, and honest limitations. The main weaknesses are mild verbosity in prose sections and a slightly non-linear workflow presentation where the quick path interrupts the numbered steps.

Suggestions

Tighten the opening paragraph and the 'SAE features are interpretable latent dimensions...' sentence, which restate concepts the rest of the skill already demonstrates.

Restructure the workflow so the numbered steps flow linearly, moving the 'Quick path (recommended)' composite-tool section before the numbered steps or clearly labeling Steps 4-5 as the optional long path up front.

Consider extracting the interpretation table and reporting-format template into a reference file to shorten the main SKILL.md overview while keeping them one level deep.

Dimension	Reasoning	Score
Conciseness	Mostly efficient and domain-specific, assuming Claude's competence; the opening paragraph and a few prose explanations (e.g. 'SAE features are interpretable latent dimensions...') could be trimmed slightly without losing value.	4 / 5
Actionability	Provides multiple copy-paste-ready, fully executable Python snippets covering the common cases (single variant composite tool, raw delta, batch saturation) with concrete parameters and expected outputs.	5 / 5
Workflow Clarity	A clear 5-step sequence with an explicit validation checkpoint (ref-residue mismatch returns an error rather than silently mutating), but the 'Quick path' interleaved between Steps 3 and 4 slightly muddies the linear sequence and the batch path lacks an explicit verify step.	4 / 5
Progressive Disclosure	Well-organized into clearly headed sections in a single file with no nested or broken references, but at ~240 lines some content (the interpretation table, reporting template) could plausibly live in reference files to keep the overview leaner.	4 / 5
	Total	17 / 20 Passed

Description

87%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 states what it does, when to use it with a concrete trigger phrase, and how it differs from neighboring skills. Minor room for improvement in trigger synonym coverage and breadth of action verbs.

Dimension	Reasoning	Score
Specificity	Names the domain and several concrete actions ('Interpret a missense variant', 'computes which interpretable SAE features ... are lost or gained') with a comprehensive list of feature categories, but the action space is somewhat narrow (interpret + compute deltas), keeping it just below a 5.	4 / 5
Completeness	Explicitly answers both 'what' (interpret a missense variant via ESMC-6B SAE feature activations) and 'when' with a concrete trigger clause ('Use when standard pathogenicity scores ... say a variant is damaging but you need a MECHANISTIC explanation').	5 / 5
Trigger Term Quality	Includes natural domain phrases users would say ('why is this variant LoF?', 'damaging', 'MECHANISTIC explanation', AlphaMissense/ClinVar), but misses some common synonyms (e.g. 'gain-of-function', 'missense mutation').	4 / 5
Distinctiveness Conflict Risk	Occupies a clear niche (SAE feature-based mechanism) and explicitly differentiates from adjacent skills ('Complements (does not replace) variant-interpretation and variant-to-mechanism skills'), giving minimal conflict risk.	5 / 5
	Total	18 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: mims-harvard/ToolUniverse
Path: plugins/tooluniverse/skills/tooluniverse-protein-sae-variant-interpretation/SKILL.md
Commit: 1b02920

Reviewed: 1 day 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.