quantummisaka/atst-tools/atst-cli

Claude Code and agent skill by quantummisaka/atst-tools

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Use when operating, testing, documenting, or extending the ATST-Tools command-line interface for ASE workflows with ABACUS/abacuslite or DeePMD-kit backends.

Install

npx vskill@latest install quantummisaka/atst-tools/atst-cli

SKILL.md

ATST CLI Skill

Use this skill for ATST-Tools CLI work in this repository.

Environment

  • Work from the repository root.
  • Prefer the project environment: conda activate atst-dev.
  • If the default Python cannot import ase, run checks with

conda run -n atst-dev <command>.

First Checks

bash
git branch --show-current
git status --short
atst --version
atst banner
atst --help
atst prepare RUN_DIR --workflow neb --init-structure INIT/STRU --final-structure FINAL/STRU -o atst_neb.yaml

Before launching expensive calculations, validate YAML:

bash
atst run --dry-run CONFIG.yaml
atst config validate CONFIG.yaml --print-normalized

Workflow Commands

Use atst run CONFIG.yaml for calculator-backed workflows:

  • neb, autoneb
  • d2s
  • dimer, sella
  • relax
  • vibration
  • irc
  • md

ABACUS workflows use calculator.name: abacus and the active abacuslite backend. DP workflows use calculator.name: dp and deepmd.calculator.DP.

On GPU hosts the optional runtime controls bind devices and thread budgets before the scientific stack is imported, and can write a run-evidence sidecar:

bash
atst run CONFIG.yaml --devices 0,1 --threads auto
atst run CONFIG.yaml --telemetry # adds runtime_evidence.json
python -m atst_tools.api.runner --config CONFIG.yaml --devices 0

--devices counts 0-based indices inside the inherited visible set (never physical renumbering); --binding round_robin maps local_rank onto a verified single-node pool for MPI launches. Without any runtime request the legacy in-process path is unchanged. See docs/user/CONFIG_REFERENCE.md section 4 and docs/user/CLI_REFERENCE.md for the full contract.

Lightweight Commands

These do not launch ABACUS or DP:

bash
atst banner
atst config validate CONFIG.yaml --print-normalized
atst abacus prepare CONFIG.yaml --structure STRUCTURE --output-dir DIR
atst abacus collect RUN_DIR --output abacus_results.json
atst neb make INIT FINAL N_IMAGES -o init_neb_chain.traj --method linear
atst neb post neb.traj --n-max N --vib-analysis
atst neb summary neb.traj --n-max N --tail 5
atst md summary md.traj --tail 5
atst md post md.traj --output-format extxyz --output-prefix md_post
atst dimer make-from-neb neb.traj --n-max N --output-traj dimer_init.traj
atst dimer summary dimer.traj --tail 5
atst relax post relax.traj --output-format traj --output restart.traj
atst relax summary relax.traj --tail 5
atst vibration post config.yaml --output vibration_results.json
atst vibration summary config.yaml
atst vibration validate vibration_results.json --structure frozen_ts.traj
atst d2s summary config.yaml --format json --output d2s_summary.json
atst traj collect frames/*.xyz -o collection.traj --no-calc
atst traj transform collection.traj --format extxyz --output-prefix collection

atst neb make --method accepts IDPP (default) or linear. summary commands are read-only monitors/post-summaries; they do not create calculators or launch ABACUS/DP. In the current 2.3.1 release, atst vibration validate is an explicit read-only screen of saved vibration data and optional frozen structure evidence. It never launches a calculator; missing evidence remains unknown.

ABACUS Wrapper Boundary

ATST-Tools is a layered wrapper:

  • It owns YAML validation, workflow orchestration, restart helpers, examples,

and common ABACUS input/output helpers.

  • It uses abacuslite as the ABACUS ASE calculator backend.
  • It does not replace ABACUS, abacuslite, Slurm, or site-specific job launchers.

Read docs/user/ABACUSLITE_WRAPPER_GUIDE.md before changing this boundary.

Validation

For CLI changes, run targeted tests first:

bash
conda run -n atst-dev pytest tests/unit/test_cli.py -q

Before reporting completion, run:

bash
conda run -n atst-dev pytest tests -q
atst --help
atst config validate examples/06_relax_H2-Au/config.yaml --print-normalized

Update docs when command behavior, config shape, or wrapper boundaries change:

  • README.md
  • docs/user/CLI_REFERENCE.md
  • docs/user/CONFIG_REFERENCE.md
  • docs/user/USER_GUIDE_CN.md
  • examples/README.md

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
54
Repository forks
16
7-day trend score
20

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/quantummisaka/atst-tools/atst-cli/badge)](https://verified-skill.com/skills/quantummisaka/atst-tools/atst-cli)

Something wrong with this skill or its page? Tell us.