菜单
Use this skill for source-grounded CP2K AIMD preparation, restart staging, generic execution handoff through cp2k_execute, and run-health inspection.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | cp2k-aimd-preparation |
| description | Use this skill for source-grounded CP2K AIMD preparation, restart staging, generic execution handoff through cp2k_execute, and run-health inspection. |
| allowed-tools | cp2k_aimd_prepare cp2k_output_summary remote_submission remote_submission_batch get_avail_remote_task md_trajectory_summary analyze_trajectory execute |
Use this skill for CP2K AIMD tasks in dynamics_worker: NVE, NVT, NPT, restart, and user-provided PLUMED metadynamics.
cp2k_aimd_prepare with recipe="nve", nvt, npt, restart, or metadynamics_user_plumed.job.inp, manifest.json, and any restart or plumed.dat files.remote_submission(task_name="cp2k_execute").cp2k_output_summary and md_trajectory_summary after outputs are collected.cp2k_aimd_preparecp2k_output_summaryremote_submissionremote_submission_batchget_avail_remote_taskmd_trajectory_summaryanalyze_trajectoryexecutesp, geo_opt, and cell_opt belong to materials_worker.timestep_fs, steps, temperature, trajectory_stride, energy_stride, and restart_stride explicitly when they matter.recipe="restart", verify the restart file and source structure before preparing the new stage.metadynamics_user_plumed, require settings.plumed_input_path; do not invent collective variables.task_name="cp2k_execute" for CP2K execution.cp2k_summary.json, job.out, CP2K energy files, trajectory files, and restart files before judging success.cp2k_output_summary for reusable run evidence: completion, return code, energy lines, SCF evidence, optimization/frequency markers, and CP2K .ener files.settings overrides from the system class and sampling objective; if that judgment remains uncertain, run a narrow literature or official documentation check before finalizing the override.settings overrides just to restate the tool baseline; only override when the user, system class, task objective, or a checked source justifies it.Return:
manifest.json pathcp2k_output_summary and trajectory summary artifact pathsreferences/cp2k_aimd_reference.mdProposal-first scientific writing pipeline. Three modes (compose/revise/hybrid) with four-layer QA pipeline. Enforces evidence-before-prose, argument-before-sections, and contracts-before-paragraphs.
Use this skill for managed MLFF molecular dynamics, restart-safe trajectory continuation, ensemble selection, and trajectory-health analysis when mlff_md is available.
Use only after a DPDispatcher-backed managed execution tool returns a failure with receipt/context fields, an ambiguous transport error, or evidence of a possible orphan job. Do not use for a pending synchronous call or ordinary success.
Use this skill before remote_submission or remote_submission_batch to build and verify canonical stage directories for registered DPDispatcher tasks, including deterministic MLFF SP/relax, MD, and NEB input layouts.
Use this skill for managed MLFF NEB optimization after a complete locally interpolated fixed-image path has been validated.
Use this skill for MLFF single-point screening, ranking, and geometry relaxation when choosing among enabled MACE, FairChem UMA, MatterSim, or ORB-v3 backends.