claude

ascg

Agentic systematic coarse-graining

4 on GitHubby Chenghao-Wuv0.6.0MITupdated 18 days ago
/

Plugin ascg from marketplace ascg-agent in Chenghao-Wu/ascg-agent. Commands read from the repo's manifest — 12 days ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (14)

  • ascg-aa-build

    Build and audit the all-atom topology and force field of a polymer, molecular liquid or mixture before any simulation (AutoPoly v2.0 OPLS-AA, a published force field where OPLS-AA lacks terms, or user files): typing errors, net charges, clashes, zero-LJ hydrogens.

  • ascg-aa-simulate

    Plan, run or import atomistic reference simulations at every approved state point (T, P) and check them before use: equilibration (polymers - RadonPy 21-step schedule; liquids - direct compression), production, engine settings, failure handling, reference QA.

  • ascg-bring-reference

    Use the user's own all-atom simulation as the reference (GROMACS, LAMMPS or any MDAnalysis-readable topology and trajectory): inspect the files, agree the state points and the mapping (atom names per residue or SMILES), build the mapped reference with ascg ref build and register it with aa import. U

  • ascg-diagnose

    Diagnose a coarse-graining task that stopped on a gate, missed convergence or shows a suspicious result, using the ascg diagnostic tools and the case library; recommend one bounded action, an escalation or a rule-change proposal.

  • ascg-fit-ibi

    Fit the CG model with multistate IBI in fastMD through ascg: submit, wait for events, read progress, and take bounded actions when a task stops or misses convergence.

  • ascg-improve

    Maintainer skill: learn from agent and operator transcripts - mine them with ascg trace mine, find repeated ad-hoc work, new failure signatures and errors, and draft tool, skill and test updates for review.

  • ascg-mapping

    Propose, check, review and approve atom-to-bead mappings from the user's description ("two beads per repeat unit", "three beads for this solvent"): resolution, bead types, end groups, bonded types; written as a SMILES mapping, checked with ascg map check, saved and approved. The mapping is always th

  • ascg-report

    Report coarse-graining results to the user: per-task summaries, study benchmark table and figure, model files, costs; plain language with numbers only from the tools.

  • ascg-setup-compute

    Set up and check compute targets for ascg (ssh or local; Slurm or plain workstations): write the user profile with ascg compute add, test it, and respect other users on shared machines.

  • ascg-setup-engine

    Install and check fastMD with tabulated and combined bending-torsion dihedrals for ascg: build plan, ctest, and the deterministic engine selfcheck.

  • ascg-specify-system

    Turn a coarse-graining request into a specification: identify the system (the user's system library or a new system), composition, state points, resolution and budget, and collect the approvals the user must give.

  • ascg-study

    Run a systematic coarse-graining study end to end with the ascg tools: specify the system, get approvals, connect compute, import or build atomistic references, review the mapping and targets, fit with multistate IBI, wait for events, diagnose, validate and report. Use for any request to coarse-grai

  • ascg-targets

    Check the target distributions of a mapped reference before fitting: RDF normalization and range, bond / angle / dihedral histograms, noise and bin widths, state coverage.

  • ascg-validate

    Interpret held-out validation of a CG model: acceptance gates, outcome classes, chain size, bonded terms, seed spread; check claims independently; handle amendment proposals.

Slash commands (2)

  • /ascg:status

    Status of an ascg study (tasks, iterations, pressure, density, outcomes, events)

  • /ascg:study

    Coarse-grain a system end to end with ascg (specify, references, mapping, fit, validate, report)

Subagents (2)

  • ascg-diagnostician

    Diagnose a coarse-graining task that stopped on a gate, missed convergence or looks suspicious, from the ascg diagnostic tools and the case library. Returns one recommendation (bounded action with parameters, escalation, or rule-change proposal) with the evidence. Use whenever an ascg task is awaiti

  • ascg-reviewer

    Independently check claims about an ascg coarse-graining result (validation outcome, report text, proposed amendment) against the tool outputs and files, before they are reported to the user.

Hooks (2)

  • PreToolUse

    python3 ${CLAUDE_PLUGIN_ROOT}/hooks/guard.py bash

  • PreToolUse

    python3 ${CLAUDE_PLUGIN_ROOT}/hooks/guard.py edit

View source on GitHub

Similar plugins

Turn raw dlt pipeline data into a Canonical Data Model: annotate sources, build an ontology, design a Kimball-style model and write @dlt.hub.transformation functions. Requires dlt[hub].

6 skills · 1 MCP server

dlt-hub/dlthub-ai-harness 65

mdclaw

claudecodex

MDClaw skills and CLI tools for preparing systems, running MD, analyzing trajectories, and packaging atomistic evidence

14 skills · hooks

matsunagalab/mdclaw 17

Generate atomistic and coarse-grained polymer systems and LAMMPS data files from SMILES

1 skill

WuGroup-XJTLU/cc-skills-ZhenghaoWu-Group 6

Maintain ascg?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex pluginslisted