Overview
ReacNetGenerator Skill is an instruction resource in the computational-chemistry-agent-skills collection, located at analysis/reacnetgenerator/SKILL.md. It guides an assistant through reactive molecular-dynamics post-processing with ReacNetGenerator and the separate reacnet-md-tools wrapper. The Skill is not the underlying analysis package or an independently running agent: its role is to organize input checks, tool selection, execution choices and output inspection.
The documented inputs include LAMMPS bond and dump trajectories, XYZ and extended XYZ files. The workflow asks for trajectory paths and atom names, with atom-name inference available from a LAMMPS data file. It also addresses scaled versus Cartesian dump coordinates, including orthorhombic and triclinic conversion through reacnet-md-tools. Periodicity decisions depend on available cell information; ambiguous atom-type mappings require clarification rather than guessed assignments.
For routine LAMMPS dump analysis, the Skill recommends rng-pipeline. It directs users to the native reacnetgenerator CLI for lower-level options, rng-query for existing .reactionabcd or .species files, and rng-webapp when a local browser interface is specifically wanted. Expected artifacts include logs, HTML/SVG/JSON reports, species and reaction files, with summary.md in the wrapper workflow and outputs organized under out/<input_basename>/.
The Skill declares uv and python3 prerequisites and notes that package resolution may need internet access unless dependencies are cached. Its HMM guidance distinguishes a quick initial pass from an explicitly requested HMM-enabled analysis. These are documented workflow recommendations, not evidence of scientific validation or tested results for a particular trajectory.
Key Features
- Guides analysis of bond, dump, xyz and extxyz reactive-MD trajectories with ReacNetGenerator.
- Routes routine runs, low-level CLI requests, existing-output queries and browser-based inspection to distinct tools.
- Describes LAMMPS coordinate conversion through reacnet-md-tools, including orthorhombic and triclinic cells.
- Supports atom-name inference from LAMMPS data files and requires clarification when the mapping is ambiguous.
- Provides decision rules for periodic boundaries, missing cell information and optional HMM treatment.
- Defines predictable output organization with run.log, report files, species/reaction artifacts and a wrapper-generated summary.
Use Cases
- Suggested evaluation: process a representative LAMMPS reactive-MD dump to assess whether coordinate handling and atom-name inference fit the simulation's file conventions.
- Inspect existing .reactionabcd and .species outputs through the documented query workflow without repeating trajectory analysis.
- Plan an XYZ or extxyz analysis by resolving cell information and periodicity before choosing the appropriate execution path.
- Suggested evaluation: compare a quick initial analysis with an explicitly requested HMM-enabled run, inspecting resulting artifacts rather than assuming either treatment is scientifically adequate.
How to Use
- Read the named Skill to understand its role within the collection and distinguish its instructions from the external analysis tools.
- Check the declared uv and python3 prerequisites. Allow for internet-dependent package resolution when packages are not cached; consult the source instructions for the supplied invocation patterns.
- Identify trajectory paths and input format. Supply atom names or an appropriate LAMMPS data file, and resolve ambiguous type mappings before analysis.
- Follow the Skill's cell and periodicity rules. Valid LAMMPS BOX BOUNDS provide cell information; XYZ files without it require a periodicity decision. Make HMM treatment explicit for the intended analysis.
- Select rng-pipeline for routine dump processing, native reacnetgenerator for required low-level options, or rng-query for existing outputs. Use the Skill's linked references rather than inventing flags.
- Inspect logs, reports and species/reaction files in the documented output location. For an initial evaluation, check input interpretation and output completeness before using the network in downstream research.