<< All versions

Skill v1.0.0

currentLLM-judged scan90/100
svd-ai-lab/sim-cli/starccm
──Details
PublishedOctober 23, 2026 at 06:43 AM
Content Hashsha256:bb4efa30b446781d...
Git SHA4e2a3c07cf05
──Files
Files (1 file, 3.7 KB)
SKILL.md3.7 KBactive
SKILL.md · 68 lines · 3.7 KB

version: "1.0.0" name: starccm-sim description: "Use when driving Simcenter STAR-CCM+ for CFD/multiphysics tasks: batch macro execution, mesh generation, solver setup, and result extraction. Choose the smallest real control path first: direct starccm+ launcher/batch, local project artifacts, native Java macros, or sim-cli when standardized runtime validation, diagnostics, sessions, and artifact handling are useful."

starccm-sim

Skill for controlling Simcenter STAR-CCM+ through the simplest real path that can produce evidence: direct starccm+ launcher/batch commands, saved .sim inspection, native Java macros, or sim-cli when its runtime layer is useful.

First principles

  1. Star-CCM+ uses Java macros, not Python scripts. Every automation script is a .java file extending StarMacro. There is no pip-installable Python SDK.
  2. Execution is batch-mode subprocess. starccm+ -batch macro.java [case.sim] — sim never imports Star-CCM+ libraries.
  3. The macro recorder is your best API discovery tool. Record GUI actions → read generated .java → adapt.
  4. License is required. CDLMD_LICENSE_FILE must point to a valid license file. The driver auto-detects license.dat near the installation.

Hard constraints

  • Script format: .java files only. Must extends StarMacro and import star.common.*.
  • Output convention: Print a JSON object via sim.println(jsonString) as the last meaningful output line. The driver's parse_output() extracts it.
  • No persistent session: Current driver is one-shot only (supports_session = False). Each sim run starts and stops Star-CCM+.
  • Startup overhead: ~15-20s per run (JVM + license checkout). Plan macros to do meaningful work per invocation.

File index

PathWhat it containsWhen to read
base/reference/java_macro_api.mdJava macro API reference — classes, methods, patternsWhen writing any macro
base/reference/batch_execution.mdCLI flags, license setup, parallel executionWhen configuring runs
base/snippets/starccm_01_smoke_test.javaMinimal connectivity testFirst run verification
base/snippets/starccm_02_pipe_flow.javaGeometry creation + volume mesh generationE2E reference
base/known_issues.mdVendor quirks and workaroundsWhen debugging failures
solver/21.02/notes.mdVersion-specific notes for 2602 (21.02)After version detection

Required protocol

  1. Direct availability probe: before requiring sim-cli, look for a real

starccm+ launcher from PATH, STARCCM_HOME/user-provided paths, common Siemens install roots, and visible STAR-CCM+ sessions. Check license signals such as CDLMD_LICENSE_FILE when present, while remembering that a path or env var is not proof of solver/license availability. For sim runtime validation, run sim check starccm from a project environment with sim-cli/plugin installed.

  1. Write macro: .java file extending StarMacro. End with sim.println(json).
  2. Lint: sim lint macro.java — checks structure before execution.
  3. Run: sim run macro.java --solver starccm — executes in batch mode.
  4. Check result: sim logs last — verify exit_code == 0 AND parsed output meets acceptance criteria.

Input classification

  • Category A (must ask user): Geometry file path, physics models, boundary condition values, acceptance criteria.
  • Category B (may default): Processor count (-np), mesh base size, iteration count.
  • Category C (derive from context): Region names, boundary names, field function names — discover via macro recorder or inspection macro.
All versions