Agent skill

Bioprobench

by PKU-YuanGroup in PKU-YuanGroup/OpenAI4S

Score an LLM's biological-protocol reasoning on the BioProBench benchmark: protocol QA, step ordering, error detection, protocol generation, and LLM-judged error reasoning; or generate the responses.

MITAuto-check passed

Install Bioprobench

skills CLI

$ npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI

$ gh skill install PKU-YuanGroup/OpenAI4S bioprobench --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy

$ git clone --depth 1 https://github.com/PKU-YuanGroup/OpenAI4S.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/bioprobench .claude/skills/bioprobench && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Paste into Claude Code

Install the "bioprobench" agent skill from https://github.com/PKU-YuanGroup/OpenAI4S/tree/main/skills/bioprobench into .claude/skills/bioprobench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "bioprobench", then confirm the skill loads.

Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

$skill-installer

$skill-installer install https://github.com/PKU-YuanGroup/OpenAI4S/tree/main/skills/bioprobench

Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.

skills CLI

$ npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install PKU-YuanGroup/OpenAI4S bioprobench --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/PKU-YuanGroup/OpenAI4S.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/bioprobench .agents/skills/bioprobench && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "bioprobench" agent skill from https://github.com/PKU-YuanGroup/OpenAI4S/tree/main/skills/bioprobench into .agents/skills/bioprobench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "bioprobench", then confirm the skill loads.

Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install PKU-YuanGroup/OpenAI4S bioprobench --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/PKU-YuanGroup/OpenAI4S.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/bioprobench .cursor/skills/bioprobench && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "bioprobench" agent skill from https://github.com/PKU-YuanGroup/OpenAI4S/tree/main/skills/bioprobench into .cursor/skills/bioprobench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "bioprobench", then confirm the skill loads.

Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

gemini skills install

$ gemini skills install https://github.com/PKU-YuanGroup/OpenAI4S.git --path skills/bioprobench

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install PKU-YuanGroup/OpenAI4S bioprobench --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/PKU-YuanGroup/OpenAI4S.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/bioprobench .gemini/skills/bioprobench && rm -rf skills-src

Use ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.

Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/

Paste into Gemini CLI

Install the "bioprobench" agent skill from https://github.com/PKU-YuanGroup/OpenAI4S/tree/main/skills/bioprobench into .gemini/skills/bioprobench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "bioprobench", then confirm the skill loads.

Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

GitHub CLI

$ gh skill install PKU-YuanGroup/OpenAI4S bioprobench

Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).

skills CLI

$ npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/PKU-YuanGroup/OpenAI4S.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/bioprobench .github/skills/bioprobench && rm -rf skills-src

Use ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.

GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/

Paste into GitHub Copilot

Install the "bioprobench" agent skill from https://github.com/PKU-YuanGroup/OpenAI4S/tree/main/skills/bioprobench into .github/skills/bioprobench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "bioprobench", then confirm the skill loads.

GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install PKU-YuanGroup/OpenAI4S bioprobench --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/PKU-YuanGroup/OpenAI4S.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/bioprobench .opencode/skills/bioprobench && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "bioprobench" agent skill from https://github.com/PKU-YuanGroup/OpenAI4S/tree/main/skills/bioprobench into .opencode/skills/bioprobench/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "bioprobench", then confirm the skill loads.

OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

Facts

Skill name: bioprobench
GitHub stars: 622
Token cost: ~2.3k tokens
SKILL.md length: 858 words
Files: 13 (incl. scripts)
Skills in repo: 17
Repo updated: Oct 9, 2026
First seen: Oct 7, 2026
Licence: MIT

At a glance

Score an LLM's biological-protocol reasoning on the BioProBench benchmark: protocol QA, step ordering, error detection, protocol generation, and LLM-judged error reasoning; or generate the responses.

SKILL.md covers The input contract is the…, Getting the data, Scoring an existing response… and Generating responses first, plus 2 more sections
Runs Python scripts from its folder; reaches github.com and api.deepseek.com

What it does

Bioprobench is an agent skill from PKU-YuanGroup/OpenAI4S. Score an LLM's biological-protocol reasoning on the BioProBench benchmark: protocol QA, step ordering, error detection, protocol generation, and LLM-judged error reasoning; or generate the responses.

Its SKILL.md is about 2.3k tokens, which your agent loads only when the skill is triggered. The skill folder holds 14 other files, including scripts (for example `README.md`, `README_zh.md` and `Scripts/LLM-as-a-judge_for_REA-ERR.py`).

The repository describes itself as: Open-source AI agent for scientific research. Analyze data in Python/R with Claude, GPT, Gemini, and more. The licence is MIT.

Example prompts

“/bioprobench”

Requirements

Python 3

What it can do on your machine

Read from SKILL.md and the folder at commit 4a72e87. It shows what the files ask for, not the result of running them.

Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Ships 1 file in scripts/ (Python), which the agent can run.
From the folder's file list and the shell code blocks in SKILL.md.
Network
Hosts in commands or code, which the agent is likely to contact:
- github.com
- api.deepseek.com
Also links to:
- huggingface.co
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Bioprobench loads about 2.3k tokens when it runs. Until then it costs about 53 tokens; SKILL.md has 858 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~53
When it runs · the whole SKILL.md, loaded when a task matches: ~2.3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from PKU-YuanGroup/OpenAI4S at commit 4a72e87, republished under its MIT licence (© PKU-YuanGroup). 858 words, ~2,317 tokens.

Download SKILL.mdSave it as .claude/skills/bioprobench/SKILL.md (or your agent's skills folder). This skill also uses 12 other files; get the full folder from GitHub.

name: bioprobench
description: Score an LLM's biological-protocol reasoning on the BioProBench benchmark: protocol QA, step ordering, error detection, protocol generation, and LLM-judged error reasoning; or generate the responses.
origin: openai4s
category: model-evaluation
metadata.display-name: BioProBench
metadata.upstream: https://github.com/YuyangSunshine/bioprotocolbench

BioProBench — protocol understanding and reasoning

Biological protocols are where a plausible-sounding answer becomes a failed experiment: a wrong dosage, a swapped step, an unflagged hazard. BioProBench scores a model on five tasks over real wet-lab protocols, roughly 5,000 instances in the full release.

Task	What it measures	Metrics returned
`PQA`	Protocol question answering — reagents, dosages, parameters	`Accuracy`, `Brier_Score`, `Failed_Rate`
`ORD`	Step ordering — reconstructing procedural sequence	`Exact_Match`, `Kendall_Tau`, `Failed_Rate`
`ERR`	Error correction — is this modified step valid	`accuracy`, `precision`, `recall`, `f1`, `failed_rate`
`GEN`	Protocol generation — synthesising steps	`BLEU`, `METEOR`, `ROUGE-L`, `KW_F1`, `Step_Recall`, `Redundancy_Penalty`, `Failed_Rate`
`REA-ERR`	Error reasoning, graded by an LLM judge	`Consistency`, `Failure_Rate`, `Total`, `Failed`, `Total_Items`, `Judged`, `Unjudged`, `Coverage`

Metric key casing differs per task — ERR returns lowercase keys, the rest are capitalised. Read them off the table above rather than guessing.

The input contract is the thing that bites

run_bioprobench_eval does not take a plain model-output file and compare it against a separate answer key. It takes one file that already has the ground truth merged into each record alongside the model's response. The upstream inference scripts produce exactly this, by adding a generated_response key to each benchmark record in place.

Hand it a file containing only model outputs and it does not raise: every record simply fails to parse and the metrics come back at zero. The envelope says so — status is "failed" when nothing scored and "partial" when some records dropped out — but still check Failed_Rate on every run. A rate of 1.0 means the input contract was violated, not that the model scored zero.

Required keys per record, per task:

Task	Model output key	Ground-truth key(s)
`PQA`	`generated_response`	`answer`
`ORD`	`generated_response`	`wrong_steps`, `correct_steps`
`ERR`	`generated_response`	`is_correct` (`true`/`false`, `1`/`0`, or `"true"`/`"false"`)
`GEN`	`generated_response`	`output` (string, or list of reference steps)
`REA-ERR`	`LLM_judge`	none — the judgment text is itself the signal

REA-ERR scores only records that carry an LLM_judge key. Consistency and Failure_Rate are over that judged subset, so read Unjudged and Coverage alongside them — a partly-judged file reports status: "partial". Populate LLM_judge with run_llm_judge_evaluation first.

Except for REA-ERR, the parser wants the answer wrapped in [ANSWER_START] … [ANSWER_END]. PQA additionally expects answer & confidence inside those tags; without the & a trailing token is read as the confidence only when it looks like one (0-100, optional %) and something is left over for the answer, otherwise the record counts as failed rather than being scored against a fabricated confidence. ORD expects a Python list literal of indices that is a genuine permutation of range(len(wrong_steps)) — anything else counts that one record as failed and leaves the rest of the run intact. Anything before a </think> or [/INST] marker is stripped first. See data/sample_pqa_output.json for two records in the exact expected shape.

Getting the data

Only the two-record sample above ships here. Download the full benchmark from Hugging Face (https://huggingface.co/BioProBench) into your working directory before running a real evaluation.

Show full SKILL.md (382 more words)Show less

Scoring an existing response file

python

from bioprobench.kernel import run_bioprobench_eval

result = run_bioprobench_eval(
    task_name="PQA",  # 'PQA' | 'ORD' | 'ERR' | 'GEN' | 'REA-ERR'
    response_file_path="/path/to/PQA_test_o3-mini.json",
)

On success the return value is:

python

{"task": "PQA", "status": "success",
 "metrics": {"Accuracy": 1.0, "Brier_Score": 0.021, "Failed_Rate": 0.0}}

status is "success" only when every record scored; "partial" when some were dropped (or, for REA-ERR, left unjudged); "failed" when none scored.

On a missing file, an unknown task name, a file that is not a non-empty JSON list of records, or any exception raised inside the evaluator, it returns a dict with no status field, so branch on "error" in result:

python

{"error": "File not found: /path/to/PQA_test_o3-mini.json"}

{"error": "out.json contains no response records, so there is nothing to"
          " score for task PQA.",
 "task": "PQA"}

{"error": "Evaluation failed during execution of task GEN on out.json:"
          " KeyError: 'output'",
 "task": "GEN", "traceback": "Traceback (most recent call last): …"}

Completion goes through host.submit_output, which takes the structured output and a required list of 1-4 completed-action bullets:

python

if "error" in result:
    raise RuntimeError(result["error"])

metrics = result["metrics"]
host.submit_output(
    result,
    [
        f"Scored the model's BioProBench {result['task']} responses.",
        f"Reported accuracy {metrics['Accuracy']:.3f} at a"
        f" {metrics['Failed_Rate']:.1%} parse-failure rate.",
    ],
)

Generating responses first

trigger_batch_inference runs the upstream inference scripts over a benchmark file and writes <TASK>_test_<model>_api.json (or ..._local.json) into the current working directory. task_name and model_name are reduced to a safe filename component first, so a slash-bearing model id lands beside the others rather than somewhere else on disk. It returns {"status": ..., "message": ...} and {"error": ...} when mode="API" is missing a key.

python

from bioprobench.kernel import trigger_batch_inference

trigger_batch_inference(
    task_name="PQA",
    test_file_path="/path/to/PQA_test.json",
    mode="API",            # "API" (OpenAI-compatible) or "Local" (HuggingFace)
    model_name="o3-mini",
    api_key="...",         # required when mode="API"
)

Both modes leave the machine: "API" sends every prompt to the configured endpoint, "Local" loads a HuggingFace checkpoint onto cuda:0. Neither is sandboxed by this skill, and the API key you pass is handed straight to the client.

Two smaller helpers:

python

from bioprobench.kernel import get_task_prompt, run_llm_judge_evaluation

# Standardised prompt for one record. Returns the prompt, or an error *string*
# beginning "Error loading prompt format:" — it does not raise.
prompt = get_task_prompt(sample, "PQA")

# Populate `LLM_judge` for REA-ERR. Returns the judgment text, or an error
# string. Requires the `openai` package and a live endpoint.
judgment = run_llm_judge_evaluation(
    sample,
    api_key="...",
    base_url="https://api.deepseek.com",
    model_name="deepseek-chat",
)

Dependencies and network access

Two things to know before the first call.

Importing kernel.py always works. Heavy dependencies are resolved per task, at call time — so ORD, ERR and REA-ERR run on a bare kernel, and a task whose dependencies are missing raises an ImportError naming exactly the packages it needs. None of them ship in the default OpenAI4S venv.

Only GEN touches the network for corpora. It fetches the NLTK tokenizer (punkt_tab, or punkt on nltk < 3.8.2) and wordnet on first use. A failure there is not fatal: the fetch is reported under a NLTK_Data_Warnings metric key. trigger_batch_inference and run_llm_judge_evaluation are the calls that send your data off-machine.

What each task needs:

Task	Runtime requirements
`ORD`, `ERR`	Standard library only
`PQA`	`numpy`, `scikit-learn`
`GEN`	`numpy`, `scikit-learn`, `nltk`, `rouge_score`, `keybert`, `sentence_transformers`, plus a first-call download of the `all-mpnet-base-v2` and `all-MiniLM-L6-v2` sentence-transformer checkpoints. Slowest task by a wide margin.
`REA-ERR`	Standard library to score; `openai` and a live endpoint to produce the judgments
`trigger_batch_inference`	`openai` for `"API"`; `transformers`, `torch`, and a GPU for `"Local"`

Citation

Adapted from the upstream BioProBench project (https://github.com/YuyangSunshine/bioprotocolbench). If you use this benchmark, cite the original work:

bibtex

@article{liu2025bioprobench,
  title={BioProBench: Comprehensive Dataset and Benchmark in Biological Protocol Understanding and Reasoning},
  author={Liu, Yuyang and Lv, Liuzhenghao and Zhang, Xiancheng and Yuan, Li and Tian, Yonghong},
  journal={ICML},
  url={https://github.com/YuyangSunshine/bioprotocolbench/tree/main},
  year={2026}
}

Files

SKILL.md and 12 other files (scripts) in skills/bioprobench of PKU-YuanGroup/OpenAI4S.

SKILL.md
README.md
README_zh.md
Scripts/LLM-as-a-judge_for_REA-ERR.py
Scripts/README.md
Scripts/README_zh.md
Scripts/generate_response.py
Scripts/generate_response_local.py
Scripts/prompt_format.py
data/README.md
data/README_zh.md
data/sample_pqa_output.json
kernel.py

Open the folder on GitHubat commit 4a72e87

Compare with similar skills

Bioprobench next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Bioprobench compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Bioprobench this skillPKU-YuanGroup/OpenAI4S	622	—	~2.3k	Automated safety check: Pass	MIT	yesterday
Benchmarkaffaan-m/ECC	277k	3 repos	~654	Automated safety check: Pass	MIT	today
Benchmarkaffaan-m/ECC	276k	—	~412	Automated safety check: Pass	MIT	yesterday
Benchmarkaffaan-m/ECC	276k	—	~330	Automated safety check: Pass	MIT	yesterday
Claw Scoreopenclaw/openclaw	392k	—	~2.5k	Automated safety check: Pass	MIT	today
Benchmark Methodologyaffaan-m/ECC	277k	1 repos	~2.5k	Automated safety check: Pass	MIT	today

Similar skills

Benchmark
affaan-m/ECC
Measure performance baselines and detect regressions across browser Core Web Vitals (LCP, INP, CLS, page weight), API endpoint latency percentiles, and build/test feedback times, with before/after…
277k GitHub starsUsed in 3 repos~654 tokens
Frontend & DesignAuto-check passed
Benchmark
affaan-m/ECC
このスキルを使用して、パフォーマンスベースラインを測定し、PR前後の回帰を検出し、スタック代替案を比較します. An agent skill from affaan-m/ECC.
276k GitHub stars~412 tokensUpdated yesterday
Auto-check passed
Benchmark
affaan-m/ECC
使用此技能测量性能基线，检测PR前后的回归，并比较堆栈替代方案。
276k GitHub stars~330 tokensUpdated yesterday
Auto-check passed
Claw Score
openclaw/openclaw
Audit or refresh OpenClaw maturity scorecard docs from root taxonomy, maturity scores, and QA evidence artifacts without using maintainer discrawl data or committed inventory reports.
392k GitHub stars~2.5k tokensUpdated today
Auto-check passed
Benchmark Methodology
affaan-m/ECC
Score a scoped competitor set into comparable profile cards: nine weighted dimensions (positioning, voice, visual craft, offer packaging, evidence, enterprise-readiness, thought leadership, pricing…
277k GitHub starsUsed in 1 repo~2.5k tokens
EducationAuto-check passed
Benchmark
androidx/androidx
Benchmarking and improving the performance of Jetpack Compose.
6.1k GitHub stars~1.1k tokensUpdated yesterday
MobileAuto-check passed

More from PKU-YuanGroup/OpenAI4S

All 17 skills in this repo

Single Cell Rna Analysis
PKU-YuanGroup/OpenAI4S
Reproducible Scanpy workflow for human or mouse 10x scRNA-seq and snRNA-seq count matrices: single-sample descriptive QC, clustering and annotation, or comparative donor-aware pseudobulk DE and Milo…
622 GitHub stars~1.3k tokensUpdated yesterday
Auto-check passed
Reaction Atom Mapping
PKU-YuanGroup/OpenAI4S
Map atoms and changed bonds for a complete reaction with RXNMapper.
622 GitHub stars~1.2k tokensUpdated yesterday
Auto-check passed
Reaction Forward Prediction
PKU-YuanGroup/OpenAI4S
Predict ranked products from reactants and reagents with ReactionT5v2-forward; use for outcome prediction or round-trip recovery.
622 GitHub stars~2k tokensUpdated yesterday
Auto-check passed
Reaction Yield Estimation
PKU-YuanGroup/OpenAI4S
Estimate yield for a fully specified reactant/reagent/product record with ReactionT5v2-yield.
622 GitHub stars~2.3k tokensUpdated yesterday
Auto-check passed
Rfdiffusion
PKU-YuanGroup/OpenAI4S
Generate de novo protein backbones with RFdiffusion for protein-target binders, hotspot-conditioned interfaces, motif scaffolding, partial diffusion, or symmetric assemblies.
622 GitHub stars~2.2k tokensUpdated yesterday
Auto-check passed
Single Step Retrosynthesis
PKU-YuanGroup/OpenAI4S
Generate ranked one-step precursor sets for a product with RetroChimera; use for disconnection ideas or expansion-policy calls.
622 GitHub stars~1.8k tokensUpdated yesterday
Auto-check passed

Questions about Bioprobench

What does Bioprobench do?

Score an LLM's biological-protocol reasoning on the BioProBench benchmark: protocol QA, step ordering, error detection, protocol generation, and LLM-judged error reasoning; or generate the responses. Bioprobench is an agent skill from PKU-YuanGroup/OpenAI4S. Score an LLM's biological-protocol reasoning on the BioProBench benchmark: protocol QA, step ordering, error detection, protocol generation, and LLM-judged error reasoning; or generate the responses.

How do I install Bioprobench in Claude Code?

Run `npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a claude-code`. Or copy the skill folder (skills/bioprobench in PKU-YuanGroup/OpenAI4S) into .claude/skills/bioprobench in your project. Claude Code loads it when a task matches its description.

How do I install Bioprobench in Codex?

Run `npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a codex`. Or copy the skill folder (skills/bioprobench in PKU-YuanGroup/OpenAI4S) into .agents/skills/bioprobench in your project. Codex loads it when a task matches its description.

Can I use Bioprobench in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add PKU-YuanGroup/OpenAI4S --skill bioprobench -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/bioprobench, .gemini/skills/bioprobench, .github/skills/bioprobench and .opencode/skills/bioprobench in your project.

What does Bioprobench need to run?

Going by SKILL.md and its folder, Bioprobench needs Python for the scripts in its folder. Our summary lists: Python 3.

Does Bioprobench access the network?

SKILL.md names 3 domains. In commands or code: github.com and api.deepseek.com; the agent is likely to contact these when it follows the instructions. As links in the text: huggingface.co. This is read from the text; nothing was executed.

Is Bioprobench safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Bioprobench use?

Bioprobench is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Bioprobench use?

About 2.3k tokens (SKILL.md is roughly 9.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Bioprobench?

Skills that share tags, products or a category with Bioprobench: Benchmark (affaan-m/ECC, 277k stars), Benchmark (affaan-m/ECC, 276k stars), Benchmark (affaan-m/ECC, 276k stars) and Claw Score (openclaw/openclaw, 392k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Bioprobench?

PKU-YuanGroup (a GitHub organization) maintains it in PKU-YuanGroup/OpenAI4S, which has 622 GitHub stars. The repository holds 17 skills in this directory. The repository was last updated on October 9, 2026.

Source: PKU-YuanGroup/OpenAI4S on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Install Bioprobench

Facts

At a glance

What it does

Example prompts

Requirements

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

BioProBench — protocol understanding and reasoning

The input contract is the thing that bites

Getting the data

Scoring an existing response file

Generating responses first

Dependencies and network access

Citation

Files

Compare with similar skills

Similar skills

More from PKU-YuanGroup/OpenAI4S

Questions about Bioprobench