esm

Comprehensive toolkit for protein language models including ESM3 (generative multimodal protein design across sequence, structure, and function) and ESM C (efficient protein embeddings and representations). Use this skill when working with protein sequences, structures, or function prediction; designing novel proteins; generating protein embeddings; performing inverse folding; or conducting protein engineering tasks. Supports both local model usage and cloud-based Forge API for scalable inference.

Quality

93%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

High

Do not use without reviewing

Quality

Content

87%

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

A dense, highly actionable body of executable code with clean one-level-deep reference navigation and no concept padding. The main weakness is the absence of explicit validation/feedback checkpoints in batch and generation workflows, which caps workflow clarity.

Suggestions

Add explicit validation/feedback steps to the batch-processing workflow (e.g., check each result for errors, retry failed generations, log successes) so the loop is validate → fix → retry rather than fire-and-forget.

For chain-of-thought generation, add a checkpoint after each step that validates the partial protein (e.g., verify the structure track produced coordinates before refining sequence, check sequence validity before predicting function).

Tighten redundancy by collapsing the standalone References section into the inline pointers already present in each capability section, and trim the Overview so it does not restate the frontmatter description.

Dimension	Reasoning	Score
Conciseness	The body is overwhelmingly executable code, model names, and configuration rather than concept padding Claude already knows; it assumes Python competence. Minor redundancy (Overview restating the description and references listed both inline and in a References section) keeps it just at 3 rather than flagging verbosity severe enough to drop to 2.	3 / 3
Actionability	Every section provides copy-paste-ready, executable Python with real imports, concrete model identifiers ('esm3-sm-open-v1', 'esmc-300m'), and real API calls; installation commands are specific. Matches the score-3 anchor.	3 / 3
Workflow Clarity	Multi-step processes are sequenced (the chain-of-thought section's Step 1/2/3) and model-selection criteria are clear, but the batch-processing section and generation workflows lack explicit validation or feedback checkpoints; per the rubric, a batch operation without validation caps this dimension at 2.	2 / 3
Progressive Disclosure	The body is an overview that signals one-level-deep references (esm3-api.md, esm-c-api.md, forge-api.md, workflows.md) — all real files verified in references/ — at the appropriate points and again in a References section, with detail appropriately split out rather than nested.	3 / 3
	Total	11 / 12 Passed

Description

100%

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, well-structured description that states concrete capabilities, provides an explicit 'Use when...' trigger clause, and occupies a clearly distinct niche using third-person voice. No fluff or over-claims are present.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'generative multimodal protein design across sequence, structure, and function', 'protein embeddings and representations', 'designing novel proteins', 'performing inverse folding', 'function prediction' — matching the score-3 anchor of multiple specific concrete actions.	3 / 3
Completeness	Explicitly answers both what ('Comprehensive toolkit for protein language models including ESM3... and ESM C...') and when via an explicit 'Use this skill when working with...' clause, satisfying the score-3 anchor.	3 / 3
Trigger Term Quality	Covers natural domain terms a user would actually say — 'protein sequences', 'structures', 'function prediction', 'designing novel proteins', 'protein embeddings', 'inverse folding', 'protein engineering' — with good breadth rather than jargon.	3 / 3
Distinctiveness Conflict Risk	A clear protein-language-model niche with distinct triggers (ESM3 generative design, ESM C embeddings, inverse folding) that is unlikely to fire for unrelated skills; not below 3 because the triggers are specific enough to avoid overlap.	3 / 3
	Total	12 / 12 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/biology/esm/SKILL.md
Commit: 52845c3

Reviewed: about 10 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.