SkillAgentSearch skills...

ProteinMPNN

Inverse-fold a protein backbone (PDB structure) into amino-acid sequence with ProteinMPNN (Dauparas et al. 2022, github.com/dauparas/ProteinMPNN).

Install / Use

npx skills add aipoch/open-science --skill proteinmpnn

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

86/100

Supported Platforms

Universal

Our assessment of ProteinMPNN

ProteinMPNN scores 86/100 on our quality scale, 2335th of 4,700 Development & Engineering skills we index (top 50%).

Its SKILL.md is 5.1 KB long, split into 5 sections with 1 code example: a solid amount of guidance for an agent.

With 4,918 GitHub stars, it is one of the more widely adopted skills in the catalogue.

Substance
26/30
Structure
15/20
Description
15/15
Adoption
16/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 2 months ago, so ProteinMPNN is actively maintained.
  • Our last check on 2026-11-03 found the source still online.
  • It is released under the Apache-2.0 license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-09-29. Automated pattern scan on 2026-09-29. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

ProteinMPNN compared with similar skills

All 4 of these similar skills score higher than ProteinMPNN; compare them before choosing.

SkillScoreStarsUpdatedFormat
ProteinMPNN (this skill)by aipoch864.9k53d agoSKILL.md
taste-skillby Leonxlnx100116.8k8d agoCLAUDE.md
Agent-Reachby Panniantong100112.3k9d agoCLAUDE.md
ai-job-searchby ai-job-search-dev10072.1ktodayCLAUDE.md
claude-howtoby luongnv8910047.3k2d agoCLAUDE.md

Frequently asked questions

How do I install ProteinMPNN?
Run npx skills add aipoch/open-science --skill ProteinMPNN. The install tabs above show the steps for each supported agent.
Which AI agents does ProteinMPNN work with?
It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
Is ProteinMPNN safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is Apache-2.0-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is ProteinMPNN still maintained?
The repository was last updated about 2 months ago, so ProteinMPNN is actively maintained.

name: proteinmpnn description: > 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. license: Apache-2.0 category: biomodels metadata: display-name: ProteinMPNN

github.com/dauparas/ProteinMPNN/blob/main/LICENSE: MIT (© 2022 Justas

Dauparas). verified 2026-06-30

third_party: - kind: weights name: ProteinMPNN license: MIT terms_url: https://github.com/dauparas/ProteinMPNN/blob/main/LICENSE

ProteinMPNN

ProteinMPNN is the default inverse-folding step in the binder pipeline: a message-passing network that sees backbone geometry only, so it is the right choice when the design surface is protein–protein and the wrong one as soon as a ligand, nucleic acid, or metal is part of the interface — ligandmpnn adds those atoms to the graph with a near-identical CLI, and solublempnn swaps in weights trained on soluble structures for an expression-biased prior. Code and weights are MIT (github.com/dauparas/ProteinMPNN). The model is small enough to run on CPU — for a handful of sequences on one backbone that is seconds and usually faster than dispatching a remote job; a GPU helps for batched campaigns (hundreds of backbones or large --num_seq_per_target). Either way the repo is cloned in-job — there is no PyPI dist and the checkpoints are bundled in the repo.

Running it

pip install torch numpy   # if not already present
git clone --depth 1 https://github.com/dauparas/ProteinMPNN.git proteinmpnn
cd proteinmpnn
python protein_mpnn_run.py \
  --pdb_path backbone.pdb --pdb_path_chains "A" \
  --out_folder out --num_seq_per_target 16 --sampling_temp "0.1"

Two flags trip almost everyone the first time. --sampling_temp is parsed as a space-separated string so one run can sweep several temperatures; a single value needs no quoting, but a multi-value sweep must be quoted ("0.1 0.2 0.3"), and commas never split — "0.1,0.2" fails the float cast. --pdb_path_chains is also space-separated inside one quoted argument ("A B"); a comma is kept as part of the chain ID.

Designs land in out/seqs/<pdb_stem>.fa. The first record is the input sequence; each design header carries score= (mean negative log-likelihood — lower is more confident), global_score=, and seq_recovery=. ProteinMPNN writes sequences only — it does not thread them back onto the backbone; if you need designed-sequence PDBs, the ligandmpnn runner writes them to backbones/ automatically and accepts --model_type protein_mpnn for the same weights.

A flat chain map in --fixed_positions_jsonl silently redesigns every residue

--fixed_positions_jsonl expects one JSON object per line keyed by the PDB stem first, then chain, then a list of 1-indexed residue numbers: {"backbone": {"A": [10, 11, 12], "B": []}}. Passing the inner {"A": [...]} directly — the obvious guess — is silently treated as "no PDB matched," and every position is redesigned. The bundled helper_scripts/make_fixed_positions_dict.py writes the correct shape from a chain and range string and is worth the extra call; the same outer-stem rule applies to --chain_id_jsonl and --tied_positions_jsonl.

Checkpoints — which one to pick

| --model_name | training noise | use | | --------------------- | -------------- | ------------------------------------------------------ | | v_48_002 | 0.02 Å | highest recovery; close-to-native redesigns | | v_48_020 (default) | 0.20 Å | de novo backbones — tolerates RFdiffusion imperfection | | v_48_030 | 0.30 Å | very rough backbones; lowest recovery | | --use_soluble_model | — | swaps to the soluble-trained set; see solublempnn |

Errors worth recognizing

| You see | It means / do this | | ---------------------------------------------------------- | ----------------------------------------------------------------------------------------------- | | KeyError: 'A' | Chain letter not in the PDB — grep '^ATOM' file.pdb \| cut -c22 \| sort -u to see what is. | | JSONDecodeError on a *_jsonl flag | The flag wants a file path, not inline JSON; write the file first. | | All positions redesigned despite --fixed_positions_jsonl | Outer PDB-stem key missing — see the gotcha above. | | ModuleNotFoundError for relative imports | Script run from the wrong cwd — cd into the cloned repo first; the imports are repo-relative. |


Next: fold the designs in complex with the target via boltz, chai1, or esmfold2 and filter on ipTM.

Related Skills

View on GitHub
GitHub Stars4.9k
CategoryDevelopment
Updated1mo ago
Forks294

Languages

TypeScript

Trust signals

100/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

No cautions