CP2K Cross-Code Input Studio

Generate, refine, explain, and cross-convert CP2K-centered input drafts for computational chemistry and materials workflows. Use when a user wants a CP2K .in...

Security Scan

The skill's code, docs, and runtime instructions are coherent with its stated purpose (CP2K draft generation and cross-code conversion), request no secrets or unrelated system access, and use only local helper scripts and reference files.

Purpose & Capability

Name/description match the provided files: reference docs, examples, and four scripts implement normalization, rendering, bundling, and conversion. No unrelated credentials, binaries, or config paths are requested.

Instruction Scope

SKILL.md instructs the agent to read the bundled reference files and use local scripts to produce normalized.json, job.inp, and report.md. The instructions focus on CP2K drafting and conversion and explicitly emphasize conservative defaults and warnings. Expected file I/O is limited to the skill folder and user-supplied structure files.

Install Mechanism

No install spec is provided; this is instruction-plus-scripts only. No downloads, external package installs, or archive extraction are requested in the metadata.

Credentials

The skill declares no required environment variables, credentials, or config paths. The behavior (reading structure files supplied by the user and writing draft files) is proportional to the stated purpose.

Persistence & Privilege

always:false and no self-modifying install steps; normal autonomous invocation is allowed (platform default) and is proportional to the skill's function.

Assessment

This skill appears coherent and self-contained, but treat outputs as draft inputs only: always review generated job.inp and report.md before running any production CP2K jobs. Pay attention to placeholder pseudopotential names and any review_required warnings (spin states, transition-metal choices, periodicity inferred from .xyz, heuristic k-points). The scripts will read any structure file you supply—do not feed sensitive or unrelated system files. If you need higher assurance, open the bundled scripts locally to inspect them before invoking the skill.

CP2K input generator

Generate a practical CP2K draft, not a fake “final validated setup”. Prefer explicit assumptions, conservative defaults, and warnings over false certainty.

Follow this workflow

1. Read references/design-rules.md.
2. Read references/job-spec-schema.md to understand the normalized contract.
3. Read only the references needed for the current task:
   • references/cp2k-task-map.md for task routing
   • references/cp2k-defaults.md for safe defaults
   • references/cp2k-kinds.md for basis/potential choices
   • references/ambiguity-policy.md for under-specified requests
   • references/structure-sources.md when the user needs a structure source rather than already having one
   • references/conversion-rules.md when the user wants Gaussian/VASP/ORCA/Quantum ESPRESSO drafts derived from the CP2K draft or normalized spec
4. Normalize the request before drafting the CP2K input.
5. Generate three deliverables whenever possible:
   • normalized.json
   • job.inp
   • report.md

Output discipline

Always make these clear:

• interpreted task
• detected system type
• periodicity assumption
• major defaults applied
• warnings / review-required items

Do not silently invent scientifically decisive inputs such as nontrivial charge, spin state for uncertain open-shell systems, or detailed periodic settings for an xyz file pretending to be a crystal.

Use the bundled scripts when helpful

These scripts are local helper utilities for deterministic text generation inside the skill folder. They are intended for transparent, offline preprocessing/rendering of CP2K draft files, not hidden network actions or privileged system changes.

• scripts/normalize-request.py: normalize a raw request JSON into the standard job spec
• scripts/render-cp2k-input.py: render a CP2K input draft from the normalized spec and a structure file
• scripts/generate_cp2k_bundle.py: one-shot pipeline that writes normalized.json, job.inp, and report.md
• scripts/convert-cp2k-input.py: convert a generated CP2K draft into a conservative Gaussian, VASP, ORCA, or Quantum ESPRESSO draft

Current supported draft space

Support is strongest for:

• molecular xyz jobs
• single-point energy
• geometry optimization
• cell optimization for periodic systems
• short test MD inputs
• vibrational analysis drafts
• conservative periodic-material drafts with heuristic k-points
• draft conversion from the generated CP2K input/spec into Gaussian, VASP, ORCA, and Quantum ESPRESSO input sets

For harder cases like transition states, NEB, unusual excited-state methods, or poorly defined transition-metal spin states, generate a conservative draft only if the workflow is still interpretable and mark the risky fields for manual review.

Do not pretend that every structure format is parsed equally deeply. If the renderer cannot deterministically recover element-wise KIND data from the provided structure file, keep the draft conservative, surface the limitation in report.md, and require manual review instead of fabricating details.

If converting a CP2K input into another code, prefer CP2K drafts generated by this skill itself. Treat arbitrary hand-written CP2K files as only partially supported unless they stay close to the emitted block structure. file, keep the draft conservative, surface the limitation in report.md, and require manual review instead of fabricating details.

If converting a CP2K input into another code, prefer CP2K drafts generated by this skill itself. Treat arbitrary hand-written CP2K files as only partially supported unless they stay close to the emitted block structure.