Use when an agent or engineer needs to drive an Ansys Fluent meshing or solver session through the sim runtime — incremental snippet execution, single-file workflows, persistent sessions via `sim connect`/`sim exec`/`sim inspect`, or one-shot script runs.
You are connected to Ansys Fluent via sim-cli. This file is the index. It tells you where to look for actual content — it does not contain the content itself.
The /connect response told you which active layers apply via:
"skills": {
"root": "<sim-skills>/fluent",
"active_sdk_layer": "0.38", // or "0.37"
"active_solver_layer":"25.2" // or null
}
Always read the index for base/, then your active sdk/<version>/,
then your active solver/<version>/. Later layers override earlier
ones on identically-named files.
Concepts, terminology, and version-agnostic control patterns:
| Path | What's there |
|---|---|
base/reference/task_templates.md | Per-workflow input templates with "Must ask if missing" lists. to identify Category A inputs before any . |
sim execbase/reference/runtime_patterns.md | The incremental "execute → inspect last.result → decide → next" loop. The canonical control protocol. |
base/reference/acceptance_checklists.md | Per-workflow acceptance criteria. Used to evaluate the final session state. |
base/reference/api_discovery.md | dir() / child_names patterns for finding renamed attributes when an API call fails. |
base/reference/tui_vs_settings.md | When to prefer the Settings API vs the legacy TUI. |
base/reference/pyfluent_overview.md | Quick orientation: what PyFluent is, what each module does. |
base/snippets/ | Numbered EX-01 step snippets (01_read_case.py through 08c_extract_all_fields.py). Built for the modern (0.38) dialect; portable to 0.37 with the material-accessor swap noted in sdk/0.37/notes.md. |
base/workflows/ | Multi-step demo scripts: meshing (watertight + fault-tolerant), solver exhaust manifold, flipchip thermal. |
Read the layer matching your active_sdk_layer field:
sdk/0.38/ — current PyFluent line.
cheat_sheet.md — most common API calls in one file.examples/ — full pyfluent.docs.pyansys.com examples (mixing_elbow,
ahmed_body, conjugate_heat_transfer, …). Each is a working Python
script you can quote from.user_guide/ — the upstream PyFluent user guide, mirrored. Search
here when an API call surprises you.sdk/0.37/ — legacy line.
notes.md — the one accessor difference vs 0.38 that bites EX-01.solver/25.2/ — Fluent 25R2 (v252) quirks (currently empty stub).These are QA artifacts for the skill itself, not content the LLM should load during a normal session:
skill_tests/ — natural-language and execution test cases used to
validate the skill end-to-end against real Fluent.tests/ — pytest integration tests.pyfluent.launch_fluent() from a snippet. sim-cli
already started the Fluent process. A second launch_fluent()
spawns a conflicting instance.ok=true and the
simulation can still be physically wrong. Validate against
base/reference/acceptance_checklists.md after the final step.After /connect succeeds: read the active task template from
base/reference/task_templates.md, gather Category A inputs from the
user, then execute incrementally per base/reference/runtime_patterns.md,
checking last.result after every step. Use snippets from
base/snippets/ (translating material-accessor calls per
sdk/<your-version>/notes.md if applicable). After the final step,
evaluate against the relevant checklist in
base/reference/acceptance_checklists.md.
For the canonical end-to-end example (the EX-01 mixing-elbow case),
the snippet sequence is 01_read_case → 02_mesh_check →
03_setup_physics → 04_setup_material → 05a_setup_bcs_ex01_ex05 →
06_hybrid_init → 07_run_150_iter → 08a_extract_outlet_temp,
followed by acceptance evaluation against the EX-01 row of
acceptance_checklists.md.