CtrlK
BlogDocsLog inGet started
Tessl Logo

chai1

Structure prediction for protein, nucleic-acid, and small-molecule complexes with the Chai-1 foundation model (Chai Discovery 2024, github.com/chaidiscovery/chai-lab). Reach for this skill to predict an antibody-antigen or protein-ligand complex from a single FASTA, to re-fold designed binders as an AlphaFold-multimer alternative, or to drive co-folding from Python for batched campaigns on a GPU.

74

Quality

93%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

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, executable reference with real code, shell equivalents, and genuinely useful operational gotchas (CHAI_DOWNLOADS_DIR, ESM VRAM cost). Its main gap is workflow clarity: a batched-operation skill would benefit from an explicit validate/rank checkpoint sequence rather than leaving the ranking thresholds as prose.

Suggestions

Add a short numbered fold→rank→filter workflow with an explicit checkpoint (e.g. 'run inference → rank by aggregate_score → keep samples with iptm > 0.5 and no clash → otherwise re-run with more recycles') to lift workflow_clarity to 3.

State the expected VRAM tier up front in the Running it section rather than burying the 80 GB recommendation only in the errors table.

Make the 'soft pass' threshold a clear pass/fail decision rule rather than advisory prose so the validation step is unambiguous.

DimensionReasoningScore

Conciseness

Lean body with executable code, a shell one-liner, and operational gotchas Claude would not already know; minimal padding and every section earns its place. It is not the level-2 'mostly efficient but could be tightened' case because no generic concepts are re-explained.

3 / 3

Actionability

Provides a complete, copy-paste-ready run_inference call with real arguments, the matching `chai-lab fold ... --use-msa-server` shell command, exact FASTA header syntax, output filenames, and ranking thresholds — fully executable guidance.

3 / 3

Workflow Clarity

The fold→rank flow is implicit and the soft-pass threshold (iptm > 0.5) and clash flags act as validation, but for a batched-campaigns skill there is no explicit numbered validate-then-proceed checkpoint or retry loop, so it caps at 2 rather than 3.

2 / 3

Progressive Disclosure

Self-contained with no bundle files and no nested references; well-organized into Running it, two gotcha sections, an error table, and a Next pointer — clean one-level structure that is easy to navigate.

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 tight, third-person description that names several concrete actions, gives an explicit 'Reach for this skill' trigger, and stakes out a distinct niche alongside Boltz/AlphaFold. It answers both what and when without padding.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — 'predict an antibody-antigen or protein-ligand complex from a single FASTA', 're-fold designed binders as an AlphaFold-multimer alternative', 'drive co-folding from Python for batched campaigns' — matching the top anchor.

3 / 3

Completeness

Explicitly answers both what ('Structure prediction for protein, nucleic-acid, and small-molecule complexes') and when ('Reach for this skill to predict...'), satisfying the top anchor with an explicit trigger clause.

3 / 3

Trigger Term Quality

Natural terms a user would say are well covered — 'protein-ligand complex', 'antibody-antigen', 'FASTA', 'AlphaFold-multimer alternative', 'co-folding', 'GPU' — going beyond jargon to common phrasings.

3 / 3

Distinctiveness Conflict Risk

Occupies a clear niche (Chai-1 co-folding vs Boltz-2/AlphaFold3) with distinct triggers; unlikely to fire for unrelated skills.

3 / 3

Total

12

/

12

Passed

Validation

81%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

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

Total

13

/

16

Passed

Repository
UnicomAI/wanwu
Reviewed

Table of Contents

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.