boltz-protein-screen

Screen existing protein binders with Boltz. Use when ranking a supplied protein, peptide, antibody, nanobody, or binder library against a target. Not for designing new proteins or screening small molecules.

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Quality

Content

85%

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

The body is highly actionable with a well-sequenced workflow that includes cost gating and retry loops, and it correctly offloads API and results detail to one-level-deep reference files. Its only weakness is conciseness: several operational instructions (background download mode, absolute paths) are repeated across multiple sections.

Suggestions

Consolidate the Claude-Code vs Codex background/non-blocking download guidance into one place (e.g., the Command Pattern or 'Always Do This') and reference it from Workflow step 6 instead of restating it three times.

Merge the repeated absolute-path / same --root-dir instructions into a single 'Always Do This' bullet rather than restating across the workflow and command comments.

Trim the longest 'Always Do This' bullets by removing restated rationale already covered in the Workflow section, keeping only the non-obvious operational constraints.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient operational detail Claude would not already know, but the Claude-Code/Codex background-download guidance and absolute-path messaging recur in Workflow step 6, the Command Pattern comments, and the 'Always Do This' bullets, so it could be tightened; not score 3 because the redundancy means not every token earns its place, and not score 1 because it avoids explaining concepts Claude already knows.	2 / 3
Actionability	Provides concrete, copy-paste-ready boltz-api commands with flags, absolute paths, payload keys, and idempotency keys, plus exact ranking metrics; not score 2 because the code/commands are fully executable rather than pseudocode or missing key details.	3 / 3
Workflow Clarity	The numbered 1–7 sequence includes an explicit estimate-cost + wait-for-confirmation gate, auth/install/sandbox retry feedback loops, and download-status completion monitoring; not score 2 because validation checkpoints and error-recovery loops are present for this batch operation rather than missing.	3 / 3
Progressive Disclosure	The body is an overview that signals one-level-deep references to real files ([references/api.md] and [references/results.md]) at the appropriate moments; not score 2 because navigation is clearly signaled and content is appropriately split rather than inlined.	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.

The description is concise, specific, and complete: it states the action, the trigger conditions with natural entity terms, and explicit negative scope that sharply limits conflict risk. It is a model 'what + when' description with no padding.

Dimension	Reasoning	Score
Specificity	Lists concrete actions ('Screen existing protein binders', 'ranking ... against a target') alongside specific entity types (protein, peptide, antibody, nanobody, binder library) and a negative scope, matching the multiple-specific-actions anchor; not score 2 because the domain and actions are concrete rather than partial.	3 / 3
Completeness	Explicitly answers both what ('Screen existing protein binders with Boltz') and when ('Use when ranking a supplied protein, peptide, antibody, nanobody, or binder library against a target') with an explicit trigger clause; not score 2 because the 'when' is stated, not merely implied.	3 / 3
Trigger Term Quality	Includes natural terms a user would say — 'protein binders', 'antibody', 'nanobody', 'screen', 'rank', 'binder library' — giving good coverage; not score 2 because common variations are present rather than missing.	3 / 3
Distinctiveness Conflict Risk	Occupies a clear niche (Boltz protein binder screening) with explicit exclusions ('Not for designing new proteins or screening small molecules'), making it unlikely to trigger for the wrong skill; not score 2 because the triggers are distinct rather than overlapping.	3 / 3
	Total	12 / 12 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: openai/plugins
Path: plugins/boltz-api-cli/skills/boltz-protein-screen/SKILL.md
Commit: 11c74d6

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.