CtrlK
BlogDocsLog inGet started
Tessl Logo

proteinmpnn

Inverse-fold a protein backbone (PDB structure) into amino-acid sequence with ProteinMPNN (Dauparas et al. 2022, github.com/dauparas/ProteinMPNN). Reach for this skill to run sequence design on RFdiffusion backbones, to redesign one chain of a PDB while holding interface residues fixed, or to generate a temperature-swept set of sequences for downstream folding.

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

Low

Low-risk findings worth noting

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.

Tight, actionable content with executable commands and well-organized sections; the main gap is the absence of an explicit validation/verification step before passing batch-designed sequences to downstream folding, which is the one place a feedback loop would matter.

Suggestions

Add a 'Verify designs' step after the run (e.g. inspect score=/seq_recovery= values and confirm the fixed-residue positions were held) before the 'Next: fold ...' pointer, to give the batch workflow an explicit checkpoint.

Note that output FASTA contains the input sequence as the first record and that ProteinMPNN does not write designed-sequence PDBs, so the reader knows to expect sequences-only output before deciding to fold.

DimensionReasoningScore

Conciseness

Lean and dense throughout — it skips explaining what proteins or inverse folding are and instead packs operational detail (flags, gotchas, checkpoint table, error table); the opening prose earns its place as decision guidance for choosing vs ligandmpnn/solublempnn rather than filler.

3 / 3

Actionability

Provides a complete, copy-paste-ready bash block (pip install, git clone, full python invocation with flags) plus concrete error-recovery commands (e.g. `grep '^ATOM' file.pdb | cut -c22 | sort -u`) and exact output paths — fully executable, not pseudocode.

3 / 3

Workflow Clarity

Steps are present (install → clone → run → inspect out/seqs) and there is a reactive error table, but for a batch operation (num_seq_per_target, temperature sweep) there is no explicit validate-the-output-before-proceeding checkpoint, which caps the score at 2 per the batch-operations guideline.

2 / 3

Progressive Disclosure

A single self-contained file with well-organized sections (Running it, fixed_positions gotcha, Checkpoints, Errors) and no nested or deep references; no bundle files are needed, so clear sectioning alone satisfies the bar.

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, third-person description that names concrete capabilities, gives an explicit use-when trigger, and carves out a distinct niche. Voice is correct (no first/second person) and it avoids fluff.

DimensionReasoningScore

Specificity

Lists several concrete actions — 'Inverse-fold a protein backbone', 'run sequence design on RFdiffusion backbones', 'redesign one chain of a PDB while holding interface residues fixed', 'generate a temperature-swept set of sequences' — matching the multiple-specific-actions anchor.

3 / 3

Completeness

Explicitly states what it does ('Inverse-fold a protein backbone ... into amino-acid sequence') and when to use it via the explicit 'Reach for this skill to ...' trigger clause, satisfying both what and when.

3 / 3

Trigger Term Quality

Covers natural domain terms a user would say — 'ProteinMPNN', 'inverse-fold', 'protein backbone', 'PDB', 'sequence design', 'RFdiffusion backbones', 'temperature-swept' — with good variation rather than jargon-only language.

3 / 3

Distinctiveness Conflict Risk

Occupies a clear niche (ProteinMPNN inverse folding on protein–protein backbones) and implicitly distinguishes itself from sibling skills like ligandmpnn/solublempnn, making wrong-skill triggering unlikely.

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.