Shaderous
NPM · SHADEROUS
GPU gas dynamics on WebGPU (hydro, MHD, gravity, AMR): load, run, measure and look at simulations
Available components
Last scanned (21 hours ago)
Scores higher than 66% of npm servers
Category Health and Science
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. How we score → Why this is hard to score →
Supply Chain Security100
- No malware found by supply-chain analysis.Pass
- No known CVEs affecting this package version or its production dependencies.Pass
- No install/post-install scripts declared.Pass
- 0 of 1 dependencies flagged as unhealthy. View diagnostics → Pass
Provenance & Transparency97
- Source repository is publicly reachable at the declared URL. View diagnostics → Pass
- Cryptographically verified build provenance (signed, bound to guillochon/shaderous). View diagnostics → Pass
- Clear OSI-approved license (MIT).Pass
- Actively maintained (last published 1 days ago).Pass
- Disclosure check failed: no security disclosure policy was found in the source repository. See how to fix → Fail
Schema Quality & AI Usability61
- AI-judged instruction clarity (good).Pass
- Context-footprint check failed: tool/resource definitions use about 3202 tokens (~139/item across 23 items; 23 tools + 0 resources), over budget; trim descriptions and params. See how to fix → Fail
- Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management13
- Stability observed for 4 of 30 days with no destabilising changes; credit accrues until the full window elapses.Partial
Tool Coverage96
- 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
- 89% of tool parameters carry a description.Partial
Tool Safety100
- No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
- We read all 23 captured tool definition(s), and no name or description among them implies an irreversible operation.Pass
- An AI judge read all 24 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
- Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.Pass
Checks we ran (16)
- Malware scan
- Known CVEs
- Install scripts
- Dependency health
- Source repository
- Provenance
- License
- Maintenance
- Security disclosure
- Schema quality
- Stability
- Tool coverage
- Injection markers
- Destructive annotations
- Judged manipulation
- MCP protocol
Does not apply to this package (2)
- Endpoint SecurityScored on hosted endpoints only. A package has no live endpoint of its own for us to probe.
- Transport & ReachabilityScored on hosted endpoints only. A package has no live endpoint of its own for us to probe.
The score does not measure everything. What the score does not measure →
Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.
- 21 Nov 26 +1
No change was recorded against any check on this day. Stability & Change Management went from 10 to 13. That category is still filling its 30-day observation window: 3 days of observed history at the previous scan, 4 at this one. The score rises as the window fills, whether or not the server changes.
- 19 Nov 26 +1
- Tool safety: pass → unverified ▼ security
- Stability: Stability not yet verified: we do not have a sandbox capture of the MCP schema this version of the package serves yet. security
- Schema quality: 2470 → 3202 ▼ functional
- Capabilities: pass → unverified ▼ functional
- Tool coverage: 100 → unverified ▼ functional
- Stability: unverified → 0.07 ▲ functional
- First check of Schema quality: unverified functional
- Package version: 0.2.0 → 0.3.0 functional
- 18 Nov 26 +15
- Malware scan: unverified → pass ▲ security
- 17 Nov 26 63
First indexed and scored.
Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.
Captured 21 Nov 2026 · Analysed npm/shaderous@0.3.0
Provenance Verified
A signed build attestation was found and verified, binding this exact artifact to the source repository it claims to come from.
| Result | Verified |
|---|---|
| Ecosystem | npm |
| Reason | Verified |
| Discovered via | Registry attestation endpoint |
| Source repo | guillochon/shaderous |
| Certificate issuer | https://token.actions.githubusercontent.com |
| Certificate SAN | https://github.com/guillochon/shaderous/.github/workflows/npm.yml@refs/tags/v0.3.0 |
| Rekor log index | 4006554366 |
| Predicate type | SLSA build provenance https://slsa.dev/provenance/v1 |
| Subject digest | sha512:256d5d286307818d5e9ede8618767c669fe0fb9ba3243a42d50ebdafe431357b7e6815d824593168009d8ca5320f7df513b12f63c796ceab2c3a31ce3 |
Background: How many MCP packages publish verified provenance →
Dependencies 1 package
| Packages resolved | 1 |
|---|---|
| Tree resolution | Complete |
Background: SBOMs and build attestations, explained →
The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →
check_setup ~162
Check a setup document before running it. Returns `ok`, an `error` if it cannot be read at all (with the key at fault), `notes` (keys dropped, values held to their range), and `issues`: each with a level (`error`: a rule broken, such as a star without isolated self-gravity, or a value that cannot work; `warning`: a choice to look at again, such as a softening under two cells, bodies that overlap or leave the box, a grid past the GPU memory), a sentence, and the path of the part or option it is about. Use it after writing or changing a document and before `load`.
| Name | Type | Req | Description |
|---|---|---|---|
| document | object | yes | The setup document (format "shaderous-setup") |
No output schema declared.
No examples provided.
control ~117
Set a panel control (or press a button, with no value) exactly as a person would, mid-run, and wait for what it starts (a shader rebuild, a restart). Examples: control("riemann", "hll"), control("eos", "gas_radiation"), control("reset"). Number fields take values in the units the panel shows; set control("units", "code") first for code units. The run stays paused.
| Name | Type | Req | Description |
|---|---|---|---|
| id | string | yes | – |
| value | string|number|boolean | – | – |
No output schema declared.
No examples provided.
controls ~40
The control panel as it stands: every control’s id, kind, label, tooltip, value and options. Use with `control` to change settings mid-run.
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
debris ~164
Debris energies, for problems that track them (describe: problems[].debris, e.g. tidal-disruption): the gas binned by its specific orbital energy about the point mass, both what has left the box (tallied as it left) and what is in it, a surviving self-bound core set apart, and the return rate of the bound debris to pericenter, dM/dt (Kepler). Returns totals, the core, the first return and the peak of dM/dt (in days and solar masses per year where the problem has units). save_path writes the full report (bins, dmdt.t and dmdt.rate) as JSON.
| Name | Type | Req | Description |
|---|---|---|---|
| save_path | string | – | Write the full report as JSON to this file |
No output schema declared.
No examples provided.
describe ~65
List the problems (presets) with their ids, dimensions, end times and parameters, and the fields, variables and choices a setup can make. Give `problem` for one problem in full, with its description.
| Name | Type | Req | Description |
|---|---|---|---|
| problem | string | – | A problem id for its full entry |
No output schema declared.
No examples provided.
diagnostics ~59
Totals and extremes read back now: mass, momentum, kinetic/internal/gravitational/total energy, max Mach number, floored cells, Jeans-unresolved cells, center of mass, point masses, and the [min, max] of every field.
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
edit ~160
Use one of the editing tools, as a person would with the pointer, applied now between steps and logged: mass (add or remove gas), velocity, temperature, obstacle (a shape in a plane of the box, cut through it along the plane normal), field (make the magnetic field stronger or weaker), pull (drag the field lines from one point to another), source, sink, star (3D), clear (the sources or the sinks), or query (the gas at a point). Positions are fractions of the box ([x, y] or [x, y, z]), velocities code units; the rest as the tool’s sliders take them. Unknown options are an error.
| Name | Type | Req | Description |
|---|---|---|---|
| name | string | yes | – |
| options | object | – | – |
No output schema declared.
No examples provided.
errors ~31
What the page has reported going wrong (uncaught errors, failed shader builds, a lost GPU device), oldest first.
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
impact ~146
What a giant impact has left (3D runs with gravity, e.g. moon-forming-impact): the gas sorted by the Kepler orbit each cell is on about the densest body into the planet (pericenter inside it), the disk around it and what escapes, with each part’s mass, the share of it that is each species and the mean dye, the disk’s angular momentum, and the mass of the moon the disk would gather into. For 3D runs with gravity (the moon-forming-impact preset). Masses in code units (Earth masses in that preset); shares are mass fractions by species, and the mean dye, which tells the two bodies apart.
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
load ~217
Build a new run from a setup and leave it paused at t = 0. A setup is a preset with overrides, or a whole setup document (`document`) that assembles a run from parts: point masses, stars, disks, gas regions, obstacles, a magnetic field. The setup uses the keys of a shareable link, as a query string ("problem=sedov&dim=2&n=512&gamma=1.4") or an object ({"problem": "sedov", "dim": 2, "n": 512}). Returns the run’s state, including `query`, the setup in force.
| Name | Type | Req | Description |
|---|---|---|---|
| document | object | – | A setup document (format "shaderous-setup", see docs/setups.md and describe().setup): the run is built from it, and `setup` then says the grid and what differs. Start from preset(id) to make one |
| setup | string|object | – | Query string, link, or object of setup keys; or a setup document itself |
No output schema declared.
No examples provided.
load_athdf ~56
Start a run from an .athdf file: one Shaderous wrote continues exactly; another code’s is loaded into the current problem. Paused at the file’s time.
| Name | Type | Req | Description |
|---|---|---|---|
| path | string | yes | The file to read |
No output schema declared.
No examples provided.
probe ~71
The gas at points given as fractions of the box ([x, y, z], z ignored in 2D): density, pressure, velocity, passive scalar, temperature (temp in code units; tempK in kelvin, with physical units), refinement level and species mass fractions.
| Name | Type | Req | Description |
|---|---|---|---|
| points | array | yes | – |
No output schema declared.
No examples provided.
record ~397
Record the run from where it is as an MP4 movie (H.264), stepping it as it goes: the current view with a color bar, x/y/z axes and the status line (grid, simulation time) over it. A frame every `every` steps, or near every `dt` of code time (at the nearest step boundary: no step is shortened, so the run is the one it would have been unrecorded). Stops at `until`, after `frames` frames, or at the end time. Set the view first. Returns the frame count, movie length and where the file went.
| Name | Type | Req | Description |
|---|---|---|---|
| axes | boolean | – | Default true |
| bitrate_mbps | number | – | Megabits per second of movie (default 0.2 bits per pixel per frame: 25 at 1080p and 60 fps) |
| colorbar | boolean | – | Default true |
| dt | number | – | A frame near every multiple of this much code time (default: a 600th of the dynamical time of the problem, so one dynamical time plays in 10 s at 60 fps) |
| every | integer | – | Or a frame every this many steps |
| fps | number | – | Playback rate (default 60) |
| frames | integer | – | Stop after this many frames |
| height | integer | – | Default 1080 (9/16 of width when width is given) |
| path | string | – | File or folder to write to (default: the current folder, under the name of the movie) |
| status | boolean | – | The status line (grid, simulation time, point masses, mesh). Default true |
| timeout_s | number | – | Give up after this much wall time, keeping what was recorded |
| until | number | – | Stop when the clock reaches this time (landed on exactly) |
| width | integer | – | Default 1920 |
No output schema declared.
No examples provided.
replay ~106
Run a session log again: load the setup it started from, step to each action and apply it, re-meshing where it re-meshed, so the run comes out as it did (to the bit, on the same machine). Returns how many actions it applied, and where it stopped if the run went another way.
| Name | Type | Req | Description |
|---|---|---|---|
| path | string | yes | The session log file |
| untilSteps | integer | – | Stop at this many steps (default: after the last action) |
No output schema declared.
No examples provided.
restart ~18
Start the run again from its initial conditions, paused.
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
run ~128
Advance the run: to time `until` (the last step is shortened to land on it exactly), by `steps` steps, or whichever comes first. Stops at the problem’s end time (stopped: "end"); to go further, load with a later `tend`. Returns t, dt, steps, steps taken, wall seconds and why it stopped.
| Name | Type | Req | Description |
|---|---|---|---|
| steps | integer | – | Take at most this many steps |
| timeout_seconds | number | – | Give up after this much wall time (default 600) |
| until | number | – | Run until the clock reads this time |
No output schema declared.
No examples provided.
save_athdf ~79
Save the state of the gas to an .athdf file (HDF5 in Athena++’s layout; yt reads it). The file carries the setup, so load_athdf continues the run exactly.
| Name | Type | Req | Description |
|---|---|---|---|
| path | string | – | File or folder to write to (default: the current folder, under the file’s own name) |
No output schema declared.
No examples provided.
screenshot ~356
A picture of the run as a PNG image. `view` changes what is drawn first: mode (volume|slice|projection), field (index, see describe), species, log, lo, hi, sliceAxis, slicePos, minRho, opacity, yaw, pitch, dist, tx, ty, tz, zoom, cx, cy, showBlocks, contours, nContours, fog, surfaceOpacity, light, box, fieldLines, fieldLineCount, mirrorImages, colormap, autoRange. Settings persist for later screenshots. `native` saves a plane at the grid's own resolution (one pixel per finest cell), with any overlays in margins beside it.
| Name | Type | Req | Description |
|---|---|---|---|
| axes | boolean | – | The axes (default false) |
| colorbar | boolean | – | The color bar (default false) |
| height | integer | – | Default: width |
| layout | string | – | 'over': the overlays over the picture's edges, as in a movie; 'beside': in margins around it, covering nothing. Default 'beside' with native, else 'over' |
| native | boolean|string | – | One pixel per cell of the finest level (the grid's own resolution, up to the largest texture): true or 'box' for the whole box, 'view' for the part in view. A plane only (a 2D run, a slice or a proje… |
| save_path | string | – | Also write the PNG to this file |
| status | boolean | – | The status line: grid, simulation time, point masses, mesh (default false) |
| view | object | – | View settings to apply first |
| width | integer | – | Default 768 |
No output schema declared.
No examples provided.
session_log ~123
The session log of the run: the setup it started from, and every edit since (the tools, settings changed, the adaptive mesh’s re-meshes), each with the step it was applied at, with the run’s diagnostics along the way. The page’s Save session log button saves the same. Returns a summary: the setup, whether it can be replayed, the actions by kind and the steps they span. save_path writes the full log as JSON, for replay.
| Name | Type | Req | Description |
|---|---|---|---|
| save_path | string | – | Write the session log as JSON to this file |
No output schema declared.
No examples provided.
slice ~284
A variable on a res x res grid on a plane through the box (perpendicular to `axis` at fraction `pos`). Returns min, max and mean, and the values (row j, column i: values[j][i]) when res is at most 64 or include_values is set; with save_path the full slice is written there instead: as a .npy file (float32, shape [res, res], for numpy’s np.load) when the name ends in .npy, else as JSON.
| Name | Type | Req | Description |
|---|---|---|---|
| axis | integer | – | Default 2 (z); always 2 in 2D |
| include_values | boolean | – | Return the values at any res (default: only when res is at most 64) |
| pos | number | – | Default 0.5 |
| res | integer | – | Default: the size of the grid, at most 256 (the MCP tool: 32) |
| save_path | string | – | Write the slice to this file: .npy (float32 for numpy), else JSON |
| variable | string | – | One of rho, press, vel1, vel2, vel3, r0, level, temp, tempK, or the name of a species. temp is the temperature in code units, tempK in kelvin (a setup with physical units). Default rho |
No output schema declared.
No examples provided.
state ~49
The run: problem, dimensions, grid, amr levels, time, timestep, steps, end time, whether it has finished, the GPU adapter, and `query`, the setup that reproduces it.
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
stats ~38
Least, greatest and volume-weighted mean of each variable over the box (rho, press, vel1-3, r0 and species mass fractions).
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
view ~104
Change what screenshots show without touching the run. Returns the view settings.
| Name | Type | Req | Description |
|---|---|---|---|
| changes | object | yes | Any of mode, field, species, log, lo, hi, sliceAxis, slicePos, minRho, opacity, yaw, pitch, dist, tx, ty, tz, zoom, cx, cy, showBlocks, contours, nContours, fog, surfaceOpacity, light, box, fieldLine… |
No output schema declared.
No examples provided.
How do I install the Shaderous MCP server?
Shaderous runs locally as an npm package, launched with npx -y shaderous mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.
npm · shaderous
claude mcp add guillochon-shaderous -- npx -y shaderous mcp
What is the Shaderous MCP server?
Shaderous is an MCP server listed in the public MCP registry as io.github.guillochon/shaderous. GPU gas dynamics on WebGPU (hydro, MHD, gravity, AMR): load, run, measure and look at simulations. This page covers its npm package (shaderous).
Is the Shaderous MCP server safe to use?
Shaderous scores 80 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 21 November 2026. It declares no install or post-install scripts. Its build provenance is signed and verified. 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 Shaderous MCP server expose?
Shaderous exposes 23 tools: describe, check_setup, load, run, state, and 18 more. Their descriptions and schemas cost roughly 2,970 tokens of context every time the server is loaded.
Is the Shaderous MCP server still maintained?
Shaderous is still listed as active in the MCP registry. We last reached this channel on 21 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 Shaderous MCP server under?
Shaderous declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.
What did VerifyMCP test on the Shaderous MCP server?
VerifyMCP ran 16 checks on Shaderous, covering Malware scan, Known CVEs, Install scripts, Dependency health and 12 more. Endpoint Security and Transport & Reachability do not apply to an npm package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.