sim CAE File Scanner
Turn existing CAE files into structured text for agents without launching a solver.
- Source repo
- svd-ai-lab/sim-cli
- Stars
- ★ 256
- Last updated
- 5d ago
- License
- Apache-2.0
- Primary language
- Python
- FA score
- 66/100 · Some gaps
At a glance
- How it runs
- Works with
- Universal · cross-platform
- Cost
- Free, no paid service needed
- Setup effort
- Low · running in minutes
- You'll need
- Typical use
- A simulation engineer needs to inventory archived Abaqus input files and extract details such as nodes, elements, materials, and analysis steps.
- Not a fit if
- Engineers who need to launch solvers or run new simulations
- Teams that need to edit CAE models in the tool
- Maintenance
- Commits in the last 30 dayslast commit 2026-11-16latest release v0.3.5 (2026-05-11)+0.6 ★/day over 7 days
- Source review
- 66/100 · Some gaps
What does this agent do, and when should you use it?
sim is a command-line tool that scans existing CAE files in batches and emits structured JSON for agents to use. It parses files through sim-cli-core without launching a solver or requiring a license server or vendor SDK. Supported formats cover COMSOL, Abaqus, Fluent, Ansys Electronics Desktop, Ansys Mechanical, Icepak Classic, Simcenter FloTHERM, Simulink, and MATLAB data files. Output is a bounded summary by default, with absolute paths redacted; use --full for complete parser results on small input sets. The repository documents installation and execution with uv, but does not describe direct integration with any named agent platform.
Run uv run sim --json scan on a CAE file or directory to identify supported formats and emit a JSON envelope with filenames, formats, summaries, and parsed data. For directory scans, --limit N caps the number of files and --no-recursive disables recursion; for an explicitly named file, --format ID forces a format. Results are bounded and absolute paths are redacted by default; --full returns complete parser results and --include-paths retains absolute paths. Run uv run sim --json describe to print the command manifest.
- A simulation engineer needs to inventory archived Abaqus input files and extract details such as nodes, elements, materials, and analysis steps.
- A thermal design team wants summaries of existing COMSOL, Fluent, or Icepak files without starting a solver.
- An engineering data developer needs to scan a directory and pass structured JSON to downstream tools.
- A team working with Simulink or MATLAB data files needs an initial view of file contents before choosing a follow-on analysis workflow.
Key setup facts
Pulled from the install and usage notes below. The project README has the current values.
- Install / run
uv add sim-cli-core
How do you install or deploy this agent?
Add sim-cli-core to an existing project with uv:
uv add sim-cli-coreHow do you use this agent?
Scan a directory and emit JSON:
uv run sim --json scan ./historical-casesScan one file and request the full result:
uv run sim --json scan ./cases/beam.inp --fullWhat are this agent's strengths and limitations?
- Reads supported existing files without launching a solver or requiring a license server or vendor SDK.
- Scans a whole directory with one command, with options to cap the file count and disable recursion.
- Bounds output and redacts absolute paths by default, which helps fit results into agent context.
- Covers engineering simulation and related data formats including Abaqus, COMSOL, Fluent, Simulink, and MATLAB.
- It reads and parses existing files; the README does not describe running solvers or editing models.
- The README claims the parser is roughly 4–20× faster than an equivalent Python script, but the provided material gives no benchmark method or data.
- The README gives a uv installation workflow but does not specify standalone installation, operating system support, or exact version requirements.
- Reported fields vary by format, so users must check the format documentation to confirm the needed data is available.
How does this agent compare with similar options?
The README compares its parser with an equivalent Python script, claiming roughly 4–20× speed, but the provided material contains no benchmark details. It also distinguishes this workflow from one that launches a solver.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| sim CAE File Scanner This agent | 66 · Some gaps | CLIFree | ★ 256 | 5d ago | Python | — |
| CAD Skills Library | 73 · Some gaps | CLIFree | ★ 27k | 1d ago | Python | Codex · Claude Code |
| REA: Reverse Engineer Anything with Agents | 84 · Good | CLIFree | ★ 471k | today | TypeScript | Codex · Claude Code · Claude.ai |
| Better Harness | 76 · Some gaps | Agent plugin / skillFree | ★ 2.5k | 10d ago | JavaScript | Codex · Claude Code |
How does FollowAgents rate this agent?
Why each dimension lost points
The server binds to localhost by default and scans redact absolute paths by default. However, /exec runs Python code, the documentation supports binding to 0.0.0.0, and the supplied server code shows no authentication control, so least privilege and external-effect controls are only partial. Skills call for inspecting inputs, verifying bounded steps, and stopping on failure, but execution itself has no user-confirmation mechanism. Data boundaries and path redaction are described; dedicated credential protections and execution rollback are not shown. Dependencies have minimum versions and a bounded simparse range, but no vulnerability-audit evidence is included. A license and NOTICE are referenced, while publisher identity is unknown and source attribution and maintenance responsibility are limited.
The code and documentation describe a stable JSON envelope, error codes, scan bounds, and plugin tiers, supporting reasonable internal consistency. However, AGENTS.md is maintenance guidance, and CLAUDE.md notes two incorrect placements of --json in agent-readability examples, so full marks are not warranted. Core dependencies have version constraints, and plugin installation paths and no-plugin capability boundaries are documented; native simparse and solver availability remain conditions. Failure responses, error codes, and diagnostic examples are clear, but this packet cannot establish equal clarity across all failure paths.
The README, shared skills, and solver skills cover defined scenarios including historical-file parsing, scripts, GUI work, and solver handoff. The documentation clearly separates standalone parsing, plugin discovery, and solver execution, and says that capability does not prove installation, licensing, or engineering correctness. Skill triggers are fairly specific, but only a sample is provided, so consistency across all plugins cannot be confirmed. Installation and configuration cover uv, standard virtual environments, and Windows examples; some GUI functions depend on a Windows interactive desktop.
The README clearly organizes formats, options, and documentation entry points; separate installation, parsing, and agent-readability documents provide useful navigation. Installation guidance covers common environments, and schema_version and the sim command show some naming stability, though the packet does not establish the full compatibility policy. Examples are practical, but no FAQ is shown. The docs describe limitations including path redaction, full-output scope, plugin requirements, and session loss on reload; a complete limitations inventory is not present. Apache-2.0 metadata agrees with LICENSE. pyproject declares a version and links a changelog, and release workflows are present, but the changelog content is not included. Repository maintenance and release steps are visible, while named ownership and update commitments are unclear.
Scan output is structured JSON with a bounded default summary, path redaction, format identifiers, and an error envelope, making it suitable for agents. Structuring CAE historical files without launching a solver is a clear marginal capability. The parser and plugin dependencies add environment costs; the stated speed comparison is only a documentation claim in this packet and receives no extra credit.
The README output example can be compared with scan and server code describing response design; source-code evidence also supports documentation about path redaction, bounded output, and error handling. Some specific claims about supported formats and speed depend on external simparse metadata or benchmark documentation not included here, limiting corroboration. Skills distinguish installation and launch evidence from engineering-result evidence, supporting separation of fact and inference, but static packet evidence cannot verify every claim.
- The HTTP server's /exec endpoint executes Python; account for the absence of authentication before exposing the service on a network interface.
- The bounded summary and default path redaction limit context and information exposure; --full and --include-paths expand the output.
- Supported-format and performance details partly rely on simparse or benchmark documentation. This review used only the supplied material and did not execute or independently verify them.
Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
Release notes excerpt
Summary - Reframed the public README around uv-first, project-local agent setup. - Added first-class Windows-safe setup wording without duplicating identical uv commands. - Updated plugin installation docs to prefer uv add ... and uv run sim .... - Removed global CLI setup guidance from the user-facing path to keep the docs agent-oriented. - Clarified agent handoff prompts to use installed solver skills without hard-coding skill paths. Package - PyPI: https://pypi.org/project/sim-cli-core/0.3.5/ - Distribution: sim-cli-core==0.3.5View on GitHub ↗ -
v0.3.0 v0.3.0 — pure plugin architecture
Release notes excerpt
sim-cli-core 0.3.0 — pure plugin architecture This release completes the multi-month migration of solver drivers out of sim-cli's source tree. The built-in driver registry is now empty — every solver driver ships as its own plugin package, discovered via [project.entry-points."sim.drivers"]. Breaking change After upgrading, sim-cli-core ships zero built-in drivers. To use any solver, install its plugin: ```sh sim plugin install <name> # e.g. ltspice, coolprop sim plugin list # see what's available + installed ``` …View on GitHub ↗
FAQ
Do I need a solver license to use it?
Does it run simulations or change models?
Will scans expose local absolute paths?
How large are the results?
Which file types are supported?
Is sim CAE File Scanner free?
Is sim CAE File Scanner safe to use?
Is sim CAE File Scanner still maintained?
Who makes sim CAE File Scanner?
Thanks — we use this to improve the page.