boltz

Structure prediction for protein, nucleic-acid, and small-molecule complexes with Boltz-2 (Passaro & Wohlwend et al. 2025, github.com/jwohlwend/boltz). Reach for this skill to validate designed binders against a target, to co-fold a protein with a SMILES or CCD ligand, or to get an open-source AlphaFold3 alternative with optional binding-affinity prediction.

Quality

89%

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

Quality

Content

96%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.

Excellent body content: dense, executable, and honest about failure modes, with concrete thresholds and an error-recovery table that make the workflow self-validating. The only structural note is that the single-file layout, while well-organized, is slightly above the size where inline-everything is unambiguously optimal.

Dimension	Reasoning	Score
Conciseness	Lean and efficient throughout — the intro stakes out positioning against sibling skills, sections deliver only Boltz-specific knowledge (MSA requirements, affinity-head semantics, kernel fallbacks), and nothing explains concepts Claude already knows.	5 / 5
Actionability	Fully executable: a complete input YAML, a copy-paste 'boltz predict' command with flags, exact output paths ('out/boltz_results_complex/predictions/complex/'), numeric decision thresholds, and an error table mapping each failure message to a fix.	5 / 5
Workflow Clarity	The flow — write YAML, run predict, read confidence JSON against explicit pass lines ('iptm > 0.5', 'complex_plddt > 0.7') — is unambiguous, and the 'Errors worth recognizing' table plus the msa/kernel sections provide explicit feedback loops for error recovery.	5 / 5
Progressive Disclosure	No bundle files exist, and the ~95-line body is well organized into clearly signaled sections (running, affinity head, MSA caveat, kernels, errors) that stay inline appropriately; it exceeds the under-50-line simple-skill threshold, so a small amount of the affinity or troubleshooting detail could arguably live in a reference file.	4 / 5
	Total	19 / 20 Passed

Description

83%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: concrete capability list, explicit use-when triggers with three distinct scenarios, and a clearly staked-out niche. Trigger-term coverage is good but not exhaustive, and the AlphaFold3-alternative framing introduces slight conflict risk with generic folding skills.

Dimension	Reasoning	Score
Specificity	Names the domain (structure prediction for protein, nucleic-acid, and small-molecule complexes) plus three concrete actions — 'validate designed binders against a target', 'co-fold a protein with a SMILES or CCD ligand', 'binding-affinity prediction' — with only minor gaps such as output formats and confidence metrics.	4 / 5
Completeness	Explicitly answers what ('Structure prediction for protein, nucleic-acid, and small-molecule complexes with Boltz-2') and when ('Reach for this skill to validate designed binders..., to co-fold a protein with a SMILES or CCD ligand, or to get an open-source AlphaFold3 alternative') with three concrete trigger phrases, matching the top anchor rather than the anchor-4 case of a less explicit 'when'.	5 / 5
Trigger Term Quality	Good natural keyword coverage ('structure prediction', 'validate designed binders', 'co-fold', 'SMILES', 'CCD', 'AlphaFold3 alternative', 'binding-affinity'), but a few common variations users would say ('protein folding', 'complex prediction') are missing.	4 / 5
Distinctiveness Conflict Risk	The Boltz-2 niche is clear and mostly distinct, but the 'open-source AlphaFold3 alternative' phrasing carries minor overlap risk with a genuine AlphaFold3 or protein-folding skill; differentiation from sibling co-folders lives in the body rather than the description.	4 / 5
	Total	17 / 20 Passed

Validation

75%

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 — 12 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 1 missing	Warning

	Total	12 / 16 Passed

Repository: aipoch/open-science
Path: resources/skills/boltz/SKILL.md
Commit: bf35648

Reviewed: 8 days 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.