# DualSPHysics MCP (pypi · dualsphysics-mcp)

Agent toolchain for the DualSPHysics SPH solver: case creation, runs, post-processing, validation.

- Trust score: 65/100 (medium)
- Change this week: −15
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-05

## Components

- pypi · `dualsphysics-mcp`: 65/100 (this document), [markdown](https://verifymcp.io/servers/vegetableno1-dualsphysics/dualsphysics-mcp.md), [page](https://verifymcp.io/servers/vegetableno1-dualsphysics/dualsphysics-mcp)

## Channel facts

- Registry: `pypi`
- Package: `dualsphysics-mcp`
- Version: `0.2.0`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-05.

- **Supply Chain Security**: 50/100
  - Malware scan not yet available for this package.
  - No known CVEs affecting this package version or its production dependencies.
  - Runs hatchling.build at install time, a recognised build step with no custom scripting around it.
  - 1 of 28 dependencies flagged as unhealthy.
- **Provenance & Transparency**: 48/100
  - Source repository is publicly reachable at the declared URL.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 41 days ago).
  - Publishes a security disclosure policy (SECURITY.md).
- **Schema Quality & AI Usability**: 61/100
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 2384 tokens (~238/item across 10 items; 10 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 100/100
  - No destabilizing schema changes in the last 30 days.
- **Tool Coverage**: 71/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 0% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 10 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 10 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a current MCP spec version (2026-07-28).

## Install

### How do I install the DualSPHysics MCP server?

DualSPHysics MCP runs locally as a PyPI package, launched with uvx dualsphysics-mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add vegetableno1-dualsphysics -- uvx dualsphysics-mcp
```

### Cursor

```json
{
  "mcpServers": {
    "vegetableno1-dualsphysics": {
      "command": "uvx",
      "args": [
        "dualsphysics-mcp"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "vegetableno1-dualsphysics": {
      "command": "uvx",
      "args": [
        "dualsphysics-mcp"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add vegetableno1-dualsphysics -- uvx dualsphysics-mcp
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "vegetableno1-dualsphysics": {
      "type": "local",
      "command": [
        "uvx",
        "dualsphysics-mcp"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add vegetableno1-dualsphysics --command uvx --arg dualsphysics-mcp
```

### Hermes

```yaml
mcp_servers:
  vegetableno1-dualsphysics:
    command: "uvx"
    args: ["dualsphysics-mcp"]
```

### Netclaw

```json
{
  "McpServers": {
    "vegetableno1-dualsphysics": {
      "Transport": "stdio",
      "Command": "uvx",
      "Arguments": [
        "dualsphysics-mcp"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add vegetableno1-dualsphysics -t stdio -c uvx -a dualsphysics-mcp
```

### Other

```json
{
  "mcpServers": {
    "vegetableno1-dualsphysics": {
      "command": "uvx",
      "args": [
        "dualsphysics-mcp"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-05 (score 65, −15)

- [security regression] Malware scan: pass → unverified

### 2026-10-29 (score 80, +15)

- [security improvement] Malware scan: unverified → pass

### 2026-10-22 (score 65, 0)

- [security] Stability: 0.97 → pass

### 2026-10-21 (score 65, +1)

No change was recorded against any check on this day. Stability & Change Management went from 93 to 97. That category is still filling its 30-day observation window: 28 days of observed history at the previous scan, 29 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-19 (score 64, +1)

No change was recorded against any check on this day. Stability & Change Management went from 87 to 90. That category is still filling its 30-day observation window: 26 days of observed history at the previous scan, 27 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-18 (score 63, −15)

- [security regression] Malware scan: pass → unverified

### 2026-10-17 (score 78, +16)

- [security improvement] Malware scan: unverified → pass

### 2026-10-15 (score 62, −14)

- [security regression] Malware scan: pass → unverified

## MCP tools (10)

### `check_environment` (~86 tokens)

Probe the DualSPHysics toolchain: paths, versions, solver features.

Reports for GenCase / CPU solver / PartVTK / MeasureTool whether each
binary was found (env var, PATH or conventional install dirs), its version,
and the solver's feature flags (a build without WaveGen cannot run wave
cases). Call this first; missing tools come with install hints.

Output parameters:

- `hint` (string)
- `jobs_dir` (string)
- `jobs_dir_exists` (boolean)
- `missing` (array)
- `omp_threads` (integer)
- `ready` (boolean): True when all four tool paths were found
- `server_version` (string)
- `tools` (array)

### `create_case` (~735 tokens)

Create a GenCase case XML (`*_Def.xml`) from a template plus overrides.

Pure Python — no solver needed. Designs a 2D dam-break-family experiment
in the XZ plane (y pinned to 0) WITHOUT hand-writing XML, with strong
parameter validation and a lattice-arithmetic particle estimate to
self-check the design before running gencase.

The `dambreak_val2d` template reproduces the official validation layout:
dp=0.01, 1 m x 2 m water column at the origin, 4 m x 3 m open-top tank,
domain (-1, 0, -1)..(4.5, 0, 3.5), TimeMax=2 s, TimeOut=0.01 s, one
vertical SWL gauge (x=0.2) and one horizontal (z=0.03). GenCase-measured
baseline counts: fluid 20,000 + bound 1,001 = 21,001.

Overrides (all optional; unknown keys are DSPH_BAD_INPUT):
\- dp: float > 0 — particle spacing (m); particle counts scale as dp^-2.
\- column_length / column_height: floats > 0 — water column at the origin.
\- tank_length / tank_height: floats > 0 — open-top tank around it.
\- obstacle: {"x": float, "width": float, "height": float} | null — solid
  box standing on the tank floor; x must be >= column_length (downstream)
  and the box must stay inside the tank. Post-processing absolute MK is 11.
\- margins: {"x_min"?, "z_min"?, "x_max"?, "z_max"?} — domain box = tank
  expanded by these margins (defaults 1.0 / 1.0 / 0.5 / 0.5 m for splash
  headroom). All must be > 0: GenCase silently clips geometry that leaves
  the domain, so the guard rejects anything that could clip.
\- domain: {"pointmin": [x, z], "pointmax": [x, z]} — explicit domain box
  (overrides margins); must strictly contain the tank.
\- time_max / time_out: floats — TimeMax / TimeOut (s); 0 < time_out <= time_max.
\- gravity, rhop0, cfl, visco: floats — physics constants (defaults
  9.81, 1000, 0.2, 0.02 artificial viscosity).
\- gauges: list of {"type": "vertical", "x": float, "z_top": float?} or
  {"type": "horizontal", "z": float, "x_end": float?} — SWL probes
  (vertical: free-surface height at fixed x; horizontal: front position
  at fixed z). Repla…

Input parameters:

- `out` (string, required)
- `overrides`
- `template` (string)

Output parameters:

- `applied` (array): Override keys applied (sorted)
- `case`
- `next_step` (string)
- `obstacle_mk_note`: Absolute MKBound id of the obstacle for post-processing
- `out_path` (string)
- `template` (string)

### `edit_case` (~233 tokens)

Edit an existing `*_Def.xml` with the create_case override vocabulary.

Reads the case (must parse as the 2D dam-break family: one fluid drawbox,
one tank drawbox, optional obstacle, vertical/horizontal SWL gauges),
applies the same overrides as create_case, re-validates the merged case
(including the domain-contains-geometry guard) and writes the result.
Files that violate the vocabulary (3D domains, multiple obstacles,
diagonal gauges, ...) are rejected with DSPH_BAD_INPUT.

Args:
    path: Case XML to edit (with or without .xml suffix).
    overrides: Same keys as create_case; e.g. {"column_height": 1.5,
        "obstacle": {"x": 2.5, "width": 0.1, "height": 0.1}}. An empty
        dict revalidates and rewrites the file in canonical form.
    save_path: Write the edited case here instead of overwriting path.

Input parameters:

- `overrides`
- `path` (string, required)
- `save_path`

Output parameters:

- `applied` (array): Override keys applied (sorted)
- `case`
- `obstacle_mk_note`
- `path` (string): Case XML that was read
- `saved_to` (string): Where the edited case was written (path or save_path)

### `describe_case` (~122 tokens)

Summarise a `*_Def.xml` case: geometry, timing, gauges, particle estimate.

Read-only self-check loop for case design: reports dp, the domain box,
water column / tank / obstacle dimensions, TimeMax/TimeOut, gauges and a
lattice-arithmetic particle estimate (fluid/bound/total) to compare with
gencase's real counts. Accepts the same 2D dam-break-family vocabulary as
edit_case (the official CaseDambreakVal2D layout parses fine).

Input parameters:

- `path` (string, required)

Output parameters:

- `case`
- `obstacle_mk_note`
- `path` (string)
- `template` (string): Vocabulary the case was recognised as

### `gencase` (~162 tokens)

Run GenCase: case XML -> Case.xml + Case.bi4 (+ preview VTK files).

Args:
    xml_path: Path to the *_Def.xml case definition (with or without .xml).
    output_dir: Output directory (default: <name>_out next to the XML).
    out_name: Output case name (default: XML stem minus _Def).
    save_modes: GenCase -save value (default "all" = bi4 + preview VTKs).
    extra_args: Extra GenCase flags (e.g. ["-dp:0.02"]).

Input parameters:

- `extra_args`
- `out_name`
- `output_dir`
- `save_modes` (string)
- `xml_path` (string, required)

Output parameters:

- `bi4_exists` (boolean)
- `command` (array)
- `duration_s` (number)
- `next_step` (string)
- `out_bi4` (string)
- `out_xml` (string)
- `particle_counts` (object): total/fluid/bound particle counts (None = not reported)
- `returncode` (integer)
- `stderr_tail` (string)
- `stdout_tail` (string)
- `vtk_files` (object)

### `run_case` (~228 tokens)

Start the CPU solver in the BACKGROUND and return a job_id.

The case must have been processed by gencase first (Case.xml + Case.bi4).
Progress is tracked with job_status; outputs land in the job directory
(Run.out, data/Part_*.bi4). Restart/cancel are not implemented; solver
flags such as -partbegin:<n> can be passed via extra_args.

Args:
    case_path: Processed case base (gencase out_xml, with/without .xml).
    dirout: Output directory (default: jobs/<job_id>).
    omp_threads: OpenMP threads (default: DSPH_OMP_THREADS or all cores).
    tmax: Override TimeMax (seconds of simulated time).
    tout: Override TimeOut (seconds between Part files).
    extra_args: Extra solver flags (e.g. ["-svres"]).

Input parameters:

- `case_path` (string, required)
- `dirout`
- `extra_args`
- `omp_threads`
- `tmax`
- `tout`

Output parameters:

- `case` (string)
- `command` (array)
- `data_dir` (string): Where Part_*.bi4 files appear
- `dirout` (string)
- `finished_at`
- `job_id` (string)
- `omp_threads` (string)
- `pid` (integer)
- `poll_hint` (string)
- `restart_note` (string)
- `returncode`
- `run_out` (string): Progress log parsed by job_status
- `started_at` (string)
- `state` (string)
- `stderr_log` (string)
- `stderr_tail` (string)
- `stdout_log` (string)

### `job_status` (~83 tokens)

Poll a background solver job: state + Run.out progress digest.

Returns the running state (plus pid/returncode), simulated time vs tmax,
percent, step counters, the solver's Time/Sec throughput (wall-clock
seconds per simulated second), the solver's own projected finish time,
particle counts and the Run.out tail.

Input parameters:

- `job_id` (string, required)

Output parameters:

- `command`
- `current_time`: Simulated time of last PART (s)
- `dirout`
- `eta`: Solver's projected finish date-time
- `eta_note` (string)
- `exception`
- `finished_at`
- `job_id` (string)
- `last_part`
- `particles`
- `percent`
- `pid`
- `poll_hint`
- `returncode`
- `run_out_exists` (boolean)
- `run_out_tail` (string)
- `started_at`
- `state` (string): running | succeeded | failed | finished
- `steps_last_part`
- `tmax`
- `total_steps`
- `wall_sec_per_sim_sec`: Solver Time/Sec column: wall-clock s per simulated s
- `wall_sec_per_sim_sec_note` (string)

### `partvtk` (~177 tokens)

PartVTK: convert Part_*.bi4 particle files to VTK for visualisation.

Args:
    job_id: Job whose data/ directory holds the Part_*.bi4 files.
    dirdata: Explicit particle data directory (alternative to job_id).
    savevtk: Output prefix (default particles/PartFluid inside the job).
    onlytype: Particle filter (default fluid only).
    variables: Variables to store, e.g. "+idp,+vel,+rhop,+press".
    first/last: Part index range.
    extra_args: Extra PartVTK flags.

Input parameters:

- `dirdata`
- `extra_args`
- `first`
- `job_id`
- `last`
- `onlytype` (string)
- `savevtk` (string)
- `variables`

Output parameters:

- `command` (array)
- `csv_files` (array)
- `duration_s` (number)
- `file_count` (integer)
- `next_step`
- `output_files` (array)
- `returncode` (integer)
- `stderr_tail` (string)
- `stdout_tail` (string)
- `vtk_files` (array)

### `measure_tool` (~307 tokens)

MeasureTool: interpolate SPH values at points -> CSV time series.

For the dam-break validation use a horizontal line of points just above
the bottom (examples/dambreak_val2d/points_damtip.txt) with density
(default -vars:-all,rhop): points beyond the surge front read 0, so the
furthest wet point is the front position that validate_dambreak compares
against the experiment.

Args:
    job_id: Job whose data/ directory holds the Part_*.bi4 files.
    dirdata: Explicit particle data directory (alternative to job_id).
    points_file: MeasureTool points file (POINTS/POINTSLIST/POINTSENDLIST).
    pointsdef: Inline points definition, e.g. "ptels[x=0:0.01:4,z=0.03]".
    variables: Variables to interpolate (default density only).
    onlytype: Particle filter (default fluid only).
    savecsv: CSV output prefix (default measure/damtip inside the job).
    savevtk: Optional VTK prefix for the measuring points.
    extra_args: Extra MeasureTool flags (a -csvsep here overrides ours).

Input parameters:

- `dirdata`
- `extra_args`
- `job_id`
- `onlytype` (string)
- `points_file`
- `pointsdef`
- `savecsv` (string)
- `savevtk`
- `variables` (string)

Output parameters:

- `command` (array)
- `csv_files` (array)
- `duration_s` (number)
- `file_count` (integer)
- `next_step`
- `output_files` (array)
- `returncode` (integer)
- `stderr_tail` (string)
- `stdout_tail` (string)
- `vtk_files` (array)

### `validate_dambreak` (~251 tokens)

Validate a 2D dam-break run: dam-tip front vs Koshizuka & Oka (1996).

Reads the CSV produced by measure_tool with the SAME points definition
(columns map to points by order), reconstructs the front position per
output time (furthest point above the wetness threshold) and reports
per-time errors plus MAE / RMSE / max error in metres and as % of the
column length.

Args:
    csv_path: MeasureTool time-history CSV.
    points: Points [[x, y, z], ...] in CSV column order.
    points_file: The points file used for the measurement.
    threshold: Wetness threshold (500 kg/m^3 for rhop; ~0.05 for vel).
    column_length: Water column length a in metres (normalisation).
    experiment: Embedded experiment (koshizuka1996).
    max_time: Ignore simulated samples beyond this time (s).

Input parameters:

- `column_length` (number)
- `csv_path` (string, required)
- `experiment` (string)
- `max_time`
- `points`
- `points_file`
- `threshold` (number)

Output parameters:

- `column_length_m` (number)
- `experiment` (string)
- `experiment_reference` (string)
- `front_series` (array)
- `impact_time_s`: When the front plateaued at the tank end (wall impact), if it did
- `mae_m` (number)
- `mae_pct_of_column` (number)
- `max_abs_error_m` (number)
- `n_samples` (integer)
- `notes`
- `rmse_m` (number)
- `threshold` (number)

## Diagnostics

Captured diagnostic sections: Provenance, Install scripts, Dependencies. The full working is on the page: https://verifymcp.io/servers/vegetableno1-dualsphysics/dualsphysics-mcp#diagnostics

## Score history

- 2026-11-05: 65
- 2026-11-04: 80
- 2026-11-03: 80
- 2026-11-02: 80
- 2026-11-01: 80
- 2026-10-31: 80
- 2026-10-30: 80
- 2026-10-29: 80
- 2026-10-28: 65
- 2026-10-27: 65
- 2026-10-26: 65
- 2026-10-25: 65
- 2026-10-24: 65
- 2026-10-23: 65
- 2026-10-22: 65
- 2026-10-21: 65
- 2026-10-20: 64
- 2026-10-19: 64
- 2026-10-18: 63
- 2026-10-17: 78
- 2026-10-16: 62
- 2026-10-15: 62
- 2026-10-14: 76
- 2026-10-13: 76
- 2026-10-12: 76
- 2026-10-08: 74
- 2026-10-05: 72
- 2026-10-04: 72
- 2026-10-03: 71
- 2026-10-02: 71

## Common questions

### What is the DualSPHysics MCP server?

DualSPHysics MCP is listed in the public MCP registry as io.github.vegetableno1/dualsphysics. Agent toolchain for the DualSPHysics SPH solver: case creation, runs, post-processing, validation. This page covers its PyPI package (dualsphysics-mcp).

### Is the DualSPHysics MCP server safe to use?

DualSPHysics MCP scores 65 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 5 November 2026. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the DualSPHysics MCP server expose?

DualSPHysics MCP exposes 10 tools: check_environment, create_case, edit_case, describe_case, gencase, and 5 more. Their descriptions and schemas cost roughly 2,384 tokens of context every time the server is loaded.

### Is the DualSPHysics MCP server still maintained?

DualSPHysics MCP is still listed as active in the MCP registry. We last reached this channel on 5 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the DualSPHysics MCP server under?

DualSPHysics MCP declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

## Links

- PyPI project: https://pypi.org/project/dualsphysics-mcp/
- Socket report: https://socket.dev/pypi/package/dualsphysics-mcp
- Repository: https://github.com/vegetableno1/DualSPHysics-mcp
- Changelog RSS feed: https://verifymcp.io/servers/vegetableno1-dualsphysics/dualsphysics-mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/vegetableno1-dualsphysics/dualsphysics-mcp.json
- HTML version of this page: https://verifymcp.io/servers/vegetableno1-dualsphysics/dualsphysics-mcp