Agent skill

Reaction Atom Mapping

by PKU-YuanGroup in PKU-YuanGroup/OpenAI4S

Map atoms and changed bonds for a complete reaction with RXNMapper.

MITAuto-check passed

Install Reaction Atom Mapping

skills CLI
$ npx skills add PKU-YuanGroup/OpenAI4S --skill reaction-atom-mapping -a claude-code

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

GitHub CLI
$ gh skill install PKU-YuanGroup/OpenAI4S reaction-atom-mapping --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/reaction-atom-mapping .claude/skills/reaction-atom-mapping && 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/

Facts

Skill name
reaction-atom-mapping
GitHub stars
622
Token cost
~1.2k tokens
SKILL.md length
528 words
Files
3
Skills in repo
17
Repo updated
First seen
Licence
MIT

At a glance

Map atoms and changed bonds for a complete reaction with RXNMapper.

  • Reactant/product correspondence and reaction-centre audits
  • SKILL.md covers Install and run, Scenario 3 benchmark contract, Derive the reaction centre and Acceptance checks, plus 2 more sections
  • Calls conda
  • Not target-only retrosynthesis

What it does

Reaction Atom Mapping is an agent skill from PKU-YuanGroup/OpenAI4S. Map atoms and changed bonds for a complete reaction with RXNMapper. Use for reactant/product correspondence and reaction-centre audits, not target-only retrosynthesis or feasibility.

Its SKILL.md is about 1.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `README.md` and `README_zh.md`).

It works with Python. 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.

When your agent uses it

  • Reactant/product correspondence and reaction-centre audits
  • Not target-only retrosynthesis

Example prompts

  • “/reaction-atom-mapping”

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

    Shell commands in SKILL.md call:

    • conda

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Links to these hosts (documentation or services it may open):

    • github.com

    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

Reaction Atom Mapping loads about 1.2k tokens when it runs. Until then it costs about 51 tokens; SKILL.md has 528 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~51
When it runs · the whole SKILL.md, loaded when a task matches
~1.2k

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); files beside SKILL.md are not scanned.

SKILL.md

The full file from PKU-YuanGroup/OpenAI4S at commit 4a72e87, republished under its MIT licence (© PKU-YuanGroup). 528 words, ~1,208 tokens.

Download SKILL.mdSave it as .claude/skills/reaction-atom-mapping/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.
name
reaction-atom-mapping
description
Map atoms and changed bonds for a complete reaction with RXNMapper. Use for reactant/product correspondence and reaction-centre audits, not target-only retrosynthesis or feasibility.
license
MIT
origin
openai4s

Reaction atom mapping

Answer one scientific question: for a reaction whose reactant and product sides are already specified, which atoms correspond across the transformation? Changed bonds derived from that mapping define a machine-readable reaction centre. This task does not propose precursors and does not establish feasibility.

Use RXNMapper, an attention-guided ALBERT mapper distributed with a local Python API and a mapping confidence value.

Install and run

Install it in the optional chemistry environment:

bash
conda create -n rxnmapper python=3.11 -y
conda run -n rxnmapper python -m pip install "rxnmapper[rdkit]==0.4.3"

Select that interpreter in its own OpenAI4S Python Cell. The switch is applied before the following Cell, so do not combine this call with the import:

python
host.env.use("rxnmapper")

After the switch succeeds, run the mapping code in a new Cell:

python
from rxnmapper import BatchedMapper

mapper = BatchedMapper(batch_size=32)
reaction = "CCBr.OCC>>CCOCC"
record = next(mapper.map_reactions_with_info([reaction]))
print(record["mapped_rxn"], record["confidence"])

Outside a Web session, where host.env.use(...) is unavailable, save the same code as a workspace script and execute it with conda run -n rxnmapper python. Do not run the bare import in the original kernel after installing into another environment.

Use BatchedMapper for campaigns because it handles invalid records without aborting the whole batch. Keep the original reaction string alongside the mapped result.

For OpenAI4S, create a manifest from the reviewed RXNMapper 0.4.3 wheel SHA in reaction_model_deployment.UPSTREAM_DISTRIBUTIONS, then call ReactionModelBackend("rxnmapper", ...) with a python_command that uses the external prefix. The foreign worker emits mapped reactions, confidence, stable atom correspondences, failures, runtime package versions, and the exact manifest fingerprint. It never downloads a model during inference.

Scenario 3 benchmark contract

Use ../retrosynthesis_planning/atom_mapping_benchmark.py for curated blind evaluation. Public inputs must be map-free. The adapter must emit the mapped reaction plus explicit stable reactant/product atom correspondences; the normalizer checks conservation, duplicate/unmapped atoms and changed bonds. Private scoring accepts pre-frozen symmetry-equivalent correspondences and excludes explicitly ambiguous reactions from whole-reaction exact accuracy, while still reporting changed-bond F1 for them.

Show full SKILL.md (244 more words)Show less

Derive the reaction centre

Parse the mapped reactant and product sides with RDKit. Build bond dictionaries keyed by sorted atom-map-number pairs, with bond type as the value. Report:

  • bonds present only on the reactant side as broken;
  • bonds present only on the product side as formed;
  • pairs present on both sides with a changed bond order;
  • mapped atoms missing from either side;
  • unmapped atoms and duplicate map numbers.

Do not silently repair unbalanced reactions, strip reagents, or move molecules between reaction fields. Those choices change the scientific object being mapped and must be recorded as preprocessing.

Acceptance checks

  • Both >> sides are present and parseable.
  • Atom-map numbers are unique within each side after mapping.
  • Every changed bond references mapped atoms.
  • Mapping confidence is preserved as the model's own score, not converted to a probability that the proposed reaction is correct.
  • For evaluation, use atom-mapping and bond-change ground truth only after the predicted mapping is fixed.

Output contract

Return original reaction SMILES, mapped reaction SMILES, mapper confidence, formed/broken/order-changed bonds, parse warnings, unmapped atoms, model version, and environment provenance.

Failure modes

SymptomAction
input contains only a productStop and use single-step-retrosynthesis; atom mapping requires both sides.
mapper returns >> or an empty recordMark the reaction invalid and preserve the original input.
confidence is lowKeep the mapping for review but do not use its reaction centre as an unquestioned label.
atom conservation failsReport imbalance separately; do not force a cosmetically balanced mapping.

Primary source: https://github.com/rxn4chemistry/rxnmapper.

© PKU-YuanGroup, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files in skills/reaction-atom-mapping of PKU-YuanGroup/OpenAI4S.

  • SKILL.md
  • README.md
  • README_zh.md

Open the folder on GitHubat commit 4a72e87

Compare with similar skills

Reaction Atom Mapping 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.

Reaction Atom Mapping compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Reaction Atom Mapping this skillPKU-YuanGroup/OpenAI4S622—~1.2kAutomated safety check: PassMIT
MCP Server Builderanthropics/skills180k63 repos~2.3kAutomated safety check: PassApache-2.0
PDF Processinganthropics/skills180k47 repos~2kAutomated safety check: PassProprietary
NotebookLM Research AssistantPleasePrompto/notebooklm-skill7.8k14 repos~2.4kAutomated safety check: NotesMIT
Manim Video Productionbrowser-use/video-use29k6 repos~3kAutomated safety check: PassMIT
PPT Masterhugohe3/ppt-master59k1 repos~2.5kAutomated safety check: PassMIT

Similar skills

  • MCP Server Builder

    anthropics/skills

    Official

    Guides the design and implementation of Model Context Protocol servers in TypeScript or Python, from tool naming and error messages to evaluation.

    180k GitHub starsUsed in 63 repos~2.3k tokens
    Agent WorkflowsAuto-check passed
  • PDF Processing

    anthropics/skills

    Official

    Handles everyday PDF jobs in Python and on the command line: extract text and tables, merge, split, rotate, watermark, fill forms, encrypt and OCR.

    180k GitHub starsUsed in 47 repos~2k tokens
    Documents & OfficeAuto-check passed
  • NotebookLM Research Assistant

    PleasePrompto/notebooklm-skill

    Lets Claude Code ask questions of your Google NotebookLM notebooks through browser automation and return answers grounded in your uploaded sources.

    7.8k GitHub starsUsed in 14 repos~2.4k tokens
    Knowledge ManagementAuto-check: notes
  • Manim Video Production

    browser-use/video-use

    Produces math and technical explainer videos with Manim Community Edition: concept animations, equation derivations, algorithm walkthroughs and data stories.

    29k GitHub starsUsed in 6 repos~3k tokens
    Media & CreativeAuto-check passed
  • PPT Master

    hugohe3/ppt-master

    Generates editable PowerPoint decks, rebuilds slides from images, fills .pptx templates and polishes existing presentations through routed workflows.

    59k GitHub starsUsed in 1 repo~2.5k tokens
    Documents & OfficeAuto-check passed
  • Scikit Learn

    zLanqing/codex-claude-academic-skills

    Machine learning in Python with scikit-learn. An agent skill from zLanqing/codex-claude-academic-skills.

    4.7k GitHub starsUsed in 16 repos~3.9k tokens
    Data & AnalyticsAuto-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
  • Bioprobench

    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.

    622 GitHub stars~2.3k 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

Works with

Questions about Reaction Atom Mapping

What does Reaction Atom Mapping do?

Map atoms and changed bonds for a complete reaction with RXNMapper. Reaction Atom Mapping is an agent skill from PKU-YuanGroup/OpenAI4S. Map atoms and changed bonds for a complete reaction with RXNMapper.

When should I use Reaction Atom Mapping?

Reaction Atom Mapping fits situations like: reactant/product correspondence and reaction-centre audits; not target-only retrosynthesis.

How do I install Reaction Atom Mapping in Claude Code?

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

How do I install Reaction Atom Mapping in Codex?

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

Can I use Reaction Atom Mapping 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 reaction-atom-mapping -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/reaction-atom-mapping, .gemini/skills/reaction-atom-mapping, .github/skills/reaction-atom-mapping and .opencode/skills/reaction-atom-mapping in your project.

What does Reaction Atom Mapping need to run?

Going by SKILL.md and its folder, Reaction Atom Mapping needs the command-line tools its instructions call (conda). Our summary lists: Python 3.

Does Reaction Atom Mapping access the network?

SKILL.md names 1 domain. As links in the text: github.com. This is read from the text; nothing was executed.

Is Reaction Atom Mapping 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. Review the folder before installing.

What licence does Reaction Atom Mapping use?

Reaction Atom Mapping is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Reaction Atom Mapping use?

About 1.2k tokens (SKILL.md is roughly 4.8k 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 Reaction Atom Mapping?

Skills that share tags, products or a category with Reaction Atom Mapping: MCP Server Builder (anthropics/skills, 180k stars), PDF Processing (anthropics/skills, 180k stars), NotebookLM Research Assistant (PleasePrompto/notebooklm-skill, 7.8k stars) and Manim Video Production (browser-use/video-use, 29k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Reaction Atom Mapping?

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.