Name: Nimbus BCI
Author: nimbusbci

Nimbus BCI MCP

AI agents build, train, and analyze BCI/EEG pipelines: data, models, experiments, live sessions.

Grade A Remote or local · Streamable HTTP · stdio API key 49 tools v0.13.2 · 14 Oct 2026

No reviews yet by nimbusbciPyPI 8,263 Runs on your machine or at its publisher’s address — not through us

93/100

trust score

downloads a day

local

or remote

383

downloads in 30d

What it does

AI agents build, train, and analyze BCI/EEG pipelines: data, models, experiments, live sessions.

Quickstart

# 1 — add it to Claude Code
claude mcp add --scope user nimbus-bci-mcp '--' uvx nimbus-mcp

# 2 — ask your agent for something these tools do
#   execution.artifacts, execution.download_artifact, pipeline.export and 46 more

Nimbus BCI costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.

PublisherVerified · since 2026

nimbusbci1 server

1 category · publishing since 2026

All 1 server

Installation

Where are you running it?

Every route below installs the same thing and ends at the same approval screen. This one can run either way — at an address its publisher runs, or as a process under your own user, with your files and your network — and this marketplace is in the path of neither.

Claude Codedeclared by this listing

This is a public server: you run it yourself and this marketplace is not in the path. You need uv (docs.astral.sh/uv): uvx comes with it. Claude Code registers it in one command; --scope user makes it available in every project. The '--' is in quotes so that PowerShell passes it on.

claude mcp add --scope user nimbus-bci-mcp '--' uvx nimbus-mcp

Reconnect, or start a new session, and the tools appear in the model’s tool list.

Or connect to its publisher’s address

Its publisher also runs it at https://nimbus-mcp.fly.dev/mcp: nothing to start on your machine, and this marketplace is not in the path.

claude mcp add --scope user --transport http nimbus-bci-mcp https://nimbus-mcp.fly.dev/mcp

Free on mcprush, and it runs from its publisher’s own package: this marketplace is not in the path, so there is no key of ours to hold and no card to add.

What it asks of you

Nothing is required, but it can take NIMBUS_API_URL (Nimbus backend URL. Defaults to the hosted Nimbus Studio (https://nimbus-studio.fly.dev); use http://127.0.0.1:8080 for a local backend or the desktop app.), NIMBUS_TOKEN (Hosted API token (nimb_…) minted in Nimbus Studio → Account → API tokens. Not needed if you already ran `nimbus-mcp login`.), NIMBUS_TOKEN_FILE (Path to a 0600 JSON file {"token": "…"} holding a hosted API token (keeps secrets out of MCP configs).), NIMBUS_MCP_KEY (Local-backend key; must match MCP_LOCAL_KEY on your Nimbus backend (dev server) or the desktop app's key.), NIMBUS_MCP_KEY_FILE (Path to a local-backend key file. The desktop app's mcp-key.json is auto-discovered without this.) if you want to set them.

What it writes

One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.

What you approve

49 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.

Tool surface

What the model actually sees. Descriptions are diffed on every release — see version history.

Only tools that write or reach the network

Trained artifacts (models/filters, e.g. *.pkl) saved by an execution.

Parameter	Type	Required	Notes
execution_id	string	yes	Run whose artifacts to list.

Download one artifact file to NIMBUS_EXPORT_DIR/executions/<id>/ and return its path.

Parameter	Type	Required	Notes
execution_id	string	yes	Run that produced the artifact.
artifact_name	string	yes	File name from execution.artifacts (e.g. "nimbus_lda.pkl").

Export the pipeline as a standalone runnable Python bundle (zip saved locally).

Parameter	Type	Required	Notes
train_graph	object	yes	Pipeline graph {nodes, connections} to export.
name	string	no	Optional name recorded inside the bundle.

Export a Nimbus dataset as a BIDS-layout zip saved to NIMBUS_EXPORT_DIR/bids/.

Parameter	Type	Required	Notes
dataset	string	yes	Dataset id — a MOABB pack id (e.g. "BNCI2014_001"), an upload path (as returned by data.upload), or a recording name. Auto-detected.
subjects	array	no	Pack exports only — subject filter (e.g. ["S01", "S03"]) to bound zip size; default exports all installed subjects.
kind	string	no	Optional explicit source kind: pack\|upload\|recording (default auto).

Export an execution's results as a BIDS-derivative zip (metrics.json, per-subject participants.tsv, protocol.json, full pipeline_snapshot.json, predictions.tsv when stored, provenance/execution.json) saved to NIMBUS_EXPORT_DIR/bids/.

Parameter	Type	Required	Notes
execution_id	string	yes	Run whose results to export (from execution.run).

Validate a pipeline graph before running.

Parameter	Type	Required	Notes
train_graph	object	yes	The graph to validate ({nodes, connections}).

Validate one node's config object against its schema (catalog.node_schema).

Parameter	Type	Required	Notes
node_type	string	yes	Node type id from catalog.nodes (e.g. "csp", "nimbus_lda").
config	object	yes	The node's config object to check.

Start a guided subject calibration session (NON-BLOCKING; confirm-gated — the device goes on a human's head). The Nimbus Studio app shows the cues on its calibration dashboard automatically; poll calibration.status.

Parameter	Type	Required	Notes
paradigm	string	no	mi \| p300 \| sart \| target_hit.
confirm	boolean	no	MUST be true — explicit user go-ahead for a session on their head.
trials_per_class	integer	no	Override the template's trial count (minimum 5).
classes	array	no	Override class list [{id,label,cue}] (MI default: left/right hand).
name	string	no	Display name recorded on the execution.
device_type	string	no	Device id from device.list (omit → template default synthetic).
connection_type	string	no	Device selector when several exist (e.g. serial vs wifi).
port	string	no	Serial/COM port for wired devices.
ip_address	string	no	Device IP for network/wifi devices.
ip_port	integer	no	Port for network devices.
stream_name	string	no	LSL stream name (LSL devices).
source_id	string	no	LSL source id.
mac_address	string	no	Bluetooth MAC (BT devices).
serial_number	string	no	Device serial (some BLE stacks).

Live snapshot of a calibration session (phase, current trial, progress, paused). Once complete, carries the recorded upload — call

Parameter	Type	Required	Notes
execution_id	string	yes	The calibration run to inspect (from calibration.start).

Pause a running calibration between trials (cues hold; resume anytime).

Parameter	Type	Required	Notes
execution_id	string	yes	The calibration run to pause.

Resume a paused calibration session.

Parameter	Type	Required	Notes
execution_id	string	yes	The calibration run to resume.

Train the subject's own classifier from a COMPLETED calibration session (NON-BLOCKING). Fetches the recorded upload, wires it into a train pipeline as a custom_data source, and starts the run. Requires a Pro plan (custom_data training is freemium-gated). The calibrate→train handoff requires a…

Parameter	Type	Required	Notes
execution_id	string	yes	The COMPLETED calibration run (from calibration.start).
paradigm	string	no	Paradigm of the recording (mi \| p300 \| sart \| target_hit); only mi has a default train template.
template_id	string	no	Train template to use (e.g. from catalog.templates); required for non-mi paradigms (mi defaults to mi_headband_csp_lda).
train_graph	object	no	Explicit train graph instead of a template; its first data node (custom_data/public_data) is rewired onto the recording.
name	string	no	Display name recorded on the training run.

Run 1-25 pipelines as ONE paced experiment (NON-BLOCKING). Returns an experimentId immediately; a background thread submits at most 2 runs at a time (min(max_concurrent, 2)), retries queue-full up to 3 times per run, and polls each execution to completion. Poll experiment.get() for per-run status…

Parameter	Type	Required	Notes
runs	array	yes	1-25 entries, each {name: str, train_graph: {nodes, connections}} (same graph shape as execution.run's train_graph).
max_concurrent	integer	no	Parallel submissions cap, clamped to 1-2 (default 2).

Experiment snapshot: status (running/completed/failed), per-run rows ({name, executionId, status, error?, metrics?}) and, once finished, aggregates {metric: {mean, std, best: {name, value}}} over completed runs only (std = population; None below 2 values).

Parameter	Type	Required	Notes
experiment_id	string	yes	The experiment to inspect (from experiment.run).

Upload an EEG file (.edf/.bdf/.mat/.csv/.txt/.tsv/.h5/.hdf5, <=500MB) to the backend and get the registered path for a custom_data node.

Parameter	Type	Required	Notes
file_path	string	yes	Local file to upload (.edf/.bdf/.mat/.csv/.txt/.tsv/.h5/.hdf5, <=500MB).
dataset_name	string	no	Optional label for the uploaded dataset.
sampling_rate	number	no	Hz for headerless CSV/TSV/TXT (REQUIRED there, e.g. 250.0).
format	string	no	Override extension-based detection (auto\|mat\|csv\|tsv\|txt\|edf\|bdf\|h5\|hdf5).

List Nimbus pipeline node types (data, preprocessing, features, models...).

Parameter	Type	Required	Notes
category	string	no	Optional filter — e.g. "data", "preprocessing", "features", "models" (exact category ids from the unfiltered list).

Full config JSON schema + input/output ports for one node type.

Parameter	Type	Required	Notes
node_type	string	yes	Node id from catalog.nodes (e.g. "csp", "nimbus_lda").

List built-in starter pipelines (MI/P300/SSVEP...).

No parameter schema on file.

Full template incl. the 'train' execGraph needed by execution.run/pipeline.validate.

Parameter	Type	Required	Notes
template_id	string	yes	Template id from catalog.templates (e.g. "mi_csp_lda").

Curated public EEG datasets (MOABB packs) available to pipelines.

Parameter	Type	Required	Notes
only_on_disk	boolean	no	Only return datasets whose data packs are present on this backend (True by default; False also lists known-but-missing sets).

Exploratory summary of a public EEG dataset (MOABB pack): channels, sampling rate, trial/class balance, per-channel µV stats, band powers and a PSD overview.

Parameter	Type	Required	Notes
dataset	string	yes	Dataset id from catalog.datasets (e.g. "BNCI2014_001").
subject	string	no	REQUIRED subject code ("S01") or comma-list cohort ("S01,S03").
mode	string	no	Which split to summarize — training \| evaluation \| all.

Exploratory summary of an EEG file (.edf/.bdf/.mat/.csv/.tsv/.txt/.h5):

Parameter	Type	Required	Notes
path	string	yes	ABSOLUTE filesystem path (~ is expanded; local backends only) or the RELATIVE upload path returned by data.upload (works on any backend).

Public benchmark leaderboard: pipeline rankings per dataset.

No parameter schema on file.

Submit a BYO python_model to a leaderboard track for server-side scoring.

Parameter	Type	Required	Notes
dataset_id	string	yes	Open dataset id (see catalog.leaderboard crossSubject datasets, e.g. "BNCI2014_001").
name	string	yes	Display name on the board (<=100 chars).
code	string	yes	Python source defining a class with fit(X, y, info) and predict(X) (<=256KB; validate first with python.validate).
class_name	string	no	Contract class to instantiate; omit to auto-detect.
requirements	array	no	Optional PyPI requirements (e.g. ["sklearn>=1.4"]).
track	string	no	"cross_subject" (default — full-data leave-one-subject-out, seed 42) or "pretrained_few_trial" (efficiency track: LOSO with a protocol-fixed 100-trials-per-subject cap; code that imports torch must de

Fetch one of your leaderboard submissions (owner-only).

Parameter	Type	Required	Notes
submission_id	string	yes	The submissionId returned by leaderboard.submit.

List your leaderboard submissions (newest first), all datasets.

No parameter schema on file.

EEG devices supported by this backend (OpenBCI, Muse, BrainBit, LSL, PiEEG...).

No parameter schema on file.

Test a device connection WITHOUT starting a stream (safe, no confirm needed).

Parameter	Type	Required	Notes
device_type	string	yes	Device id from device.list (e.g. "brainbit", "muse").
connection_type	string	no	Device-specific selector when several exist (e.g. serial vs wifi).
port	string	no	Serial/COM port for wired devices.
ip_address	string	no	Device IP for network/wifi devices.
ip_port	integer	no	Port for network devices.
stream_name	string	no	LSL stream name (LSL devices).
source_id	string	no	LSL source id.
mac_address	string	no	Bluetooth MAC (BT devices).
serial_number	string	no	Device serial (some BLE stacks).

Connect an EEG device and START a live streaming session on the user's head.

Parameter	Type	Required	Notes
device_type	string	yes	Device id from device.list (e.g. "brainbit").
confirm	boolean	no	MUST be true to start — the explicit user go-ahead for a live session on their head; anything else is refused with zero requests.
session_id	string	no	Optional existing session to resume/reuse.
chunk_size	integer	no	Samples per streamed chunk (default 125).
n_channels	integer	no	Channel count to open (default 8).
connection_type	string	no	Device-specific selector (e.g. serial vs wifi).
port	string	no	Serial/COM port for wired devices.
ip_address	string	no	Device IP for network/wifi devices.
ip_port	integer	no	Port for network devices.
stream_name	string	no	LSL stream name (LSL devices).
source_id	string	no	LSL source id.
mac_address	string	no	Bluetooth MAC (BT devices).
serial_number	string	no	Device serial (some BLE stacks).
idle_timeout_sec	integer	no	Watchdog: stop+disconnect after this many seconds without a status poll (default 900; 0 disables).

Live snapshot of a streaming session (running, deviceConnected).

Parameter	Type	Required	Notes
session_id	string	yes	The streaming session to inspect.

Stop a streaming session and disconnect the device (always safe to call).

Parameter	Type	Required	Notes
session_id	string	yes	The streaming session to stop.

Confusion-matrix figure for one execution: heatmap with trial counts + per-class accuracy bars, rendered as a PNG the assistant can see. Use it right after execution.results to eyeball WHICH classes are confused, not just the accuracy number.

Parameter	Type	Required	Notes
execution_id	string	yes	Run to plot (from execution.run / execution.results).
matrix	string	no	"eval" (held-out, default) or "train" (fit-set sanity check).
save	boolean	no	Also write the PNG to NIMBUS_EXPORT_DIR/plots/ and return its path.

Figure over an EEG data source — the visual companion to

Parameter	Type	Required	Notes
kind	string	yes	Figure type: psd \| topomap \| bandpower \| spectrogram.
source	string	yes	"dataset" (MOABB pack), "upload" (path from data.upload), or "local" (absolute path, local backend only).
dataset	string	no	MOABB dataset id, e.g. "BNCI2014_001" (source=dataset).
subject	string	no	Subject id, e.g. "S01" (source=dataset).
mode	string	no	Pack split to load: training \| evaluation \| all (source=dataset).
path	string	no	Upload path (source=upload) or absolute local path (source=local).
band	string	no	Band for kind=topomap: delta\|theta\|alpha\|beta\|gamma (default alpha).
channel_idx	integer	no	0-based channel for kind=spectrogram (default 0).
save	boolean	no	Also write the PNG to NIMBUS_EXPORT_DIR/plots/ and return its path.

ERP butterfly figure from epoched data: one subplot per class, all channel means (thin) + across-channel mean (bold), time-locked to the trial event. Shows whether class-conditioned evoked structure exists BEFORE training anything.

Parameter	Type	Required	Notes
source	string	yes	"dataset" (MOABB pack), "upload" (path from data.upload), or "local" (absolute path, local backend only).
dataset	string	no	MOABB dataset id, e.g. "BNCI2014_001" (source=dataset).
subject	string	no	Subject id, e.g. "S01" (source=dataset).
mode	string	no	Pack split to load: training \| evaluation \| all (source=dataset).
path	string	no	Upload path (source=upload) or absolute local path (source=local).
max_channels	integer	no	Channels drawn per class subplot (default 16, max 64).
save	boolean	no	Also write the PNG to NIMBUS_EXPORT_DIR/plots/ and return its path.

Leaderboard bar chart for one dataset: accuracy with 95% CI

Parameter	Type	Required	Notes
dataset_id	string	yes	Dataset to chart, e.g. "BNCI2014_001" (see catalog.leaderboard).
track	string	no	"within_session" (curated 5-fold), "crossSubject" (LOSO, curated + BYO submissions), or "pretrained_few_trial" (the efficiency track: LOSO at the protocol's fixed per-subject trial budget — the figure
save	boolean	no	Also write the PNG to NIMBUS_EXPORT_DIR/plots/ and return its path.

Create a project (container for one pipeline document). Returns projectId.

Parameter	Type	Required	Notes
name	string	yes	Project display name (e.g. "MI CSP-LDA sweep").
description	string	no	Optional longer description shown in the studio.

List projects owned by the current principal (agent work included).

No parameter schema on file.

Save a pipeline graph into a project (visible on the studio canvas).

Parameter	Type	Required	Notes
project_id	string	yes	Target project (from project.create/project.list).
train_graph	object	yes	Pipeline graph {nodes, connections} to persist.
name	string	no	Optional pipeline name stored on the document.
subject	string	no	Optional dataset subject code (e.g. "S01") for this pipeline.

Load a project's saved pipeline (train graph + meta) for editing/re-running.

Parameter	Type	Required	Notes
project_id	string	yes	Project whose pipeline document to load.

Statically check a python_model script (BYO model class) without running it.

Parameter	Type	Required	Notes
code	string	yes	Python source of the model class (<=256KB UTF-8).
class_name	string	no	Contract class to instantiate; omit to auto-detect the first class defining fit+predict.

Start a pipeline run (NON-BLOCKING). Returns executionId — poll with execution.get() until status is completed/failed, then execution.results().

Parameter	Type	Required	Notes
train_graph	object	yes	Pipeline graph {nodes: [{id, type, config}], connections: [{from, to}]} as built by catalog.template/pipeline.validate.
name	string	no	Display name for the run (shown in the Runs list).
description	string	no	Optional longer description of the experiment.
subject	string	no	Optional dataset subject code (e.g. "S01") recorded with the run.
layout	object	no	Optional canvas positions {nodes: {id: {x, y}}}; a deterministic grid is synthesized when omitted.

Cancel a running execution.

Parameter	Type	Required	Notes
execution_id	string	yes	The run to terminate (from execution.run/execution.list).

Execution status summary (status: running/completed/failed/cancelled).

Parameter	Type	Required	Notes
execution_id	string	yes	The run to check (from execution.run/execution.list).

Recent executions. Optional status filter (running/completed/failed/cancelled).

Parameter	Type	Required	Notes
limit	integer	no	Maximum number of runs to return (default 20).
status	string	no	Filter by run status: running/completed/failed/cancelled.

Metrics for a completed run. Trimmed by default (accuracy, kappa, ITR, confusion matrix, per-class); full=True returns the complete result object.

Parameter	Type	Required	Notes
execution_id	string	yes	The completed run to fetch metrics for.
full	boolean	no	Return the backend's complete result object (all fields).

Cluster-aware group statistics across executions: pools the

Parameter	Type	Required	Notes
execution_ids	array	yes	1–25 executions (from execution.run) to pool. Runs without stored group stats (fewer than 3 groups) are reported with nulls but do not fail the request.
chance	number	no	Chance level in [0, 1]; default 1/nClasses per execution.

Live snapshot of a streaming session: latest prediction + recent window, signal quality (meanChannelQuality, snrDb, artifactProbability), indicators, running stats. Poll this while a session runs. Works with OR without a deployed model: modelless hardware streams (no classifier) report live signal…

Parameter	Type	Required	Notes
session_id	string	yes	The streaming session to read telemetry for.
window	integer	no	How many recent predictions/chunks to include (default 50).

Who you are authenticated as: account email, plan (isPro / pioneer), this month's free-run quota, and — with a hosted token — the token name and days until it expires. Call this first when setup guidance appears or to check which credential a session uses.

No parameter schema on file.

Free

$0 forever

Every tool, no call limit
No card, and mcprush charges nothing for it
Source is public; its licence metadata reads NOASSERTION rather than naming a licence
Runs on your machine — nothing of it reaches our gateway
Nothing to cap, because mcprush bills nothing

Nothing here is billable: mcprush charges nothing to install Nimbus BCI or to call it, at any volume.

ATrust score 93 / 100

Release 0.13.2 · scanned 14 Oct 2026

Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.

Threat score100

Client score93

Findings0 threat · 1 capability

Capabilityhigh

Coveragesource

Verificationrepo

Surface digest43538839b2b8f921a7d09817c126b4486f70f462a129a3244572d1e759dd844b

Score vectorthreat 100 → client 93

Term	Level	What it prices	Points
capability-exposure	high	capability blast radius (high) — client exposure if the model is manipulated	−6
verification-discount	repo	publisher verification (public source) — no provenance, but the source is public and inspectable	−1
coverage-honesty	source	inspection depth (source) — how much of the target the scan could see	−0

score = clamp(round(threatScore − E_cap − E_ver − E_cov), 0, 100). No model, no opinion, no network call at scoring time: the same methodology version against the same bytes always produces this number, and every line above is a field of the report rather than a summary of it.

What the scan could actually read

A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.

Tool surfaceread

Implementation sourceread

Package metadataread

Live transportnot available

Tools were statically extracted from the published source (49 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command "npx -y <package>".

Capability — what it could do if the model were manipulated

Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. high is the level these add up to.

Tool	Capability tags	Why the tag was assigned
execution.artifacts	no tags
execution.download_artifact	untrusted-input	untrusted-input: keyword "download"
pipeline.export	no tags
bids.export_dataset	sensitive-source	sensitive-source: keyword "export_data"
bids.export_execution	no tags
pipeline.validate	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/validate-pipeline") at src/nimbus_mcp/tools/build.py:23
pipeline.validate_node	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/validate-node-config") at src/nimbus_mcp/tools/build.py:33
calibration.start	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/execute") at src/nimbus_mcp/tools/calibration.py:229
calibration.status	no tags
calibration.pause	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/pause-calibration") at src/nimbus_mcp/tools/calibration.py:277
calibration.resume	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/resume-calibration") at src/nimbus_mcp/tools/calibration.py:287
calibration.train	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/execute") at src/nimbus_mcp/tools/calibration.py:399
experiment.run	no tags
experiment.get	no tags
data.upload	no tags
catalog.nodes	no tags
catalog.node_schema	no tags
catalog.templates	no tags
catalog.template	no tags
catalog.datasets	no tags
foundation.compare	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/foundation/compare") at src/nimbus_mcp/tools/foundation.py:51
data.inspect_dataset	no tags
data.inspect_file	no tags
catalog.leaderboard	no tags
leaderboard.submit	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/leaderboard/submissions") at src/nimbus_mcp/tools/leaderboard.py:76
leaderboard.status	untrusted-input	untrusted-input: keyword "fetch"
leaderboard.mine	no tags
device.list	no tags
device.test	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/deployment/test-device") at src/nimbus_mcp/tools/live.py:121
stream.start	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/hardware/connect") at src/nimbus_mcp/tools/live.py:189, handler sends an HTTP POST ( client.post("/api/hardware/start-stream") at src/nimbus_mcp/tools/live.py:192, handler sends an HTTP POST ( client.post("/api/hardware/disconnect") at src/nimbus_mcp/tools/live.py:205
stream.status	no tags
stream.stop	no tags
plots.confusion	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/plots/confusion") at src/nimbus_mcp/tools/plots.py:91
plots.dataset	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/plots/dataset") at src/nimbus_mcp/tools/plots.py:131
plots.erp	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/plots/erp") at src/nimbus_mcp/tools/plots.py:172
plots.leaderboard	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/plots/leaderboard") at src/nimbus_mcp/tools/plots.py:209
project.create	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/projects") at src/nimbus_mcp/tools/projects.py:27
project.list	no tags
project.save	no tags
project.load	no tags
python.validate	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/python-model/validate") at src/nimbus_mcp/tools/python.py:64
execution.run	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/execute") at src/nimbus_mcp/tools/run.py:103
execution.cancel	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/cancel") at src/nimbus_mcp/tools/run.py:119
execution.get	no tags
execution.list	no tags
execution.results	untrusted-input	untrusted-input: keyword "fetch"
stats.grouped	external-sink	external-sink: handler sends an HTTP POST ( client.post("/api/stats/grouped") at src/nimbus_mcp/tools/stats.py:49
stream.telemetry	no tags
account.whoami	no tags

Toxic-flow graph

The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.

execution.download_artifact → bids.export_dataset → calibration.pause

A cross-tool exfiltration chain exists: execution.download_artifact → bids.export_dataset → calibration.pause.

Supply chain and provenance

Integrityfirst-seen

Surface digest43538839b2b8f921

Scanned2026-10-14

This is the first scan of this surface here, so there is nothing yet to compare it against.

Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.

Nimbus BCI MCP

What it does

Quickstart

Where are you running it?

Tool surface

What counts against your monthly calls

What the scan could actually read

Capability — what it could do if the model were manipulated

Toxic-flow graph

Supply chain and provenance

Release history

See also