tools: add physics-sanity + Docker MCP scaffolding + tools/README

- tools/physics-sanity/ — JS library of cross-node balance helpers
  (mass / hydraulic / hydraulic-power / oxygen-transfer / energy) with
  7 unit tests + a CLI demo. Designed for `require()` from per-node
  integration tests where shape-based unit tests miss physically-
  impossible plant states.
- tools/docker-compose.yml + tools/mcp/{node-red-admin,influxdb,browser}
  scaffolding — placeholder Dockerfiles + a ROADMAP.md for the Node-RED
  admin MCP. Compose file is the target shape for the Q3-2026 migration
  to the central MCP server; the per-service Dockerfile stays in this
  repo as the canonical definition either way. Implementations are TODO.
- tools/README.md — top-level tooling index; documents the CI order for
  running every tool on a PR.
- .gitignore: ignore tools/.env (developer-specific MCP endpoints).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
znetsixe
2026-05-19 10:16:47 +02:00
parent edef1cecbf
commit 6e6699c763
13 changed files with 548 additions and 0 deletions

57
tools/README.md Normal file
View File

@@ -0,0 +1,57 @@
# EVOLV tooling (`tools/`)
Repo-local tools and MCP services for EVOLV development. All Node.js
native (no Python toolchain). Each tool encodes a rule that we've
previously discovered through a bug; skipping them re-opens those bugs.
See `CLAUDE.md` § "Tooling (Docker-first, local now, central later)"
for the operating doctrine.
## Tools (CLI)
| Tool | Path | What it does | Run |
|---|---|---|---|
| `contract-verify` | `tools/contract-verify/` | Diffs `nodes/<n>/CONTRACT.md` topic table vs `src/commands/index.js` | `node tools/contract-verify/bin/contract-verify.js` |
| `flow-lint` | `tools/flow-lint/` | Lints `examples/*.flow.json` against the flow-layout rule | `node tools/flow-lint/bin/flow-lint.js` |
| `wiki-gen` | `tools/wiki-gen/` | Regenerates AUTOGEN topic-contract blocks in per-node wikis | `node tools/wiki-gen/bin/wiki-gen.js` |
| `output-manifest-verify` | `tools/output-manifest-verify/` | Enforces the output-coverage manifest rule | `node tools/output-manifest-verify/bin/output-manifest-verify.js` |
| `physics-sanity` | `tools/physics-sanity/` | Library of cross-node balance helpers; import from tests | `require('../../tools/physics-sanity')` |
CI-friendly: every tool accepts `--json` (JSON output) and exits non-zero
on findings.
## MCP services (Docker)
See `tools/mcp/README.md`. Three services scaffolded:
| Service | Purpose | Status |
|---|---|---|
| `mcp-node-red-admin` | Wraps Node-RED admin HTTP API | TODO (scaffold) |
| `mcp-influxdb` | Telemetry query + assertion | TODO (scaffold) |
| `mcp-browser` | Headless Playwright against the dashboard | TODO (scaffold) |
Start (once impls land):
```bash
cd tools && docker compose --profile mcp up -d
```
## CI integration
Recommended order on every PR:
```bash
node tools/contract-verify/bin/contract-verify.js # 1. CONTRACT vs registry
node tools/flow-lint/bin/flow-lint.js # 2. Flow JSON shapes
node tools/wiki-gen/bin/wiki-gen.js --check # 3. Wiki AUTOGEN blocks
node tools/output-manifest-verify/bin/output-manifest-verify.js # 4. Manifest coverage
```
Each is fast (<1 s on the whole repo).
## Adding a new tool
1. `tools/<name>/package.json` with a `bin` entry.
2. `tools/<name>/bin/<name>.js` — must accept `--json` and exit 1 on drift.
3. `tools/<name>/README.md` — one-page docs.
4. Add a row to this README + a row to `CLAUDE.md` § Tooling.
5. Wire into the CI snippet above.

46
tools/docker-compose.yml Normal file
View File

@@ -0,0 +1,46 @@
name: evolv-tools
services:
# ----------------------------------------------------------------------
# Local MCP services for the Claude Code tooling stack.
# Migration note (2026-05): these are deliberately local. When the
# central MCP server comes online (target Q3 2026), each service moves
# to the shared infra; this compose file becomes the spec for what to
# provision there. The build context + Dockerfile in tools/mcp/<name>/
# stays in this repo as the canonical definition.
# ----------------------------------------------------------------------
mcp-node-red-admin:
build:
context: ./mcp/node-red-admin
container_name: evolv-mcp-node-red-admin
environment:
NODE_RED_HOST: ${NODE_RED_HOST:-http://host.docker.internal:1880}
NODE_RED_TOKEN: ${NODE_RED_TOKEN:-}
stdin_open: true
tty: true
profiles: ["mcp"]
mcp-influxdb:
build:
context: ./mcp/influxdb
container_name: evolv-mcp-influxdb
environment:
INFLUX_URL: ${INFLUX_URL:-http://host.docker.internal:8086}
INFLUX_TOKEN: ${INFLUX_TOKEN:-}
INFLUX_ORG: ${INFLUX_ORG:-wbd}
INFLUX_BUCKET: ${INFLUX_BUCKET:-telemetry}
stdin_open: true
tty: true
profiles: ["mcp"]
mcp-browser:
build:
context: ./mcp/browser
container_name: evolv-mcp-browser
environment:
DASHBOARD_URL: ${DASHBOARD_URL:-http://host.docker.internal:1880/dashboard}
HEADLESS: ${HEADLESS:-true}
stdin_open: true
tty: true
profiles: ["mcp"]

97
tools/mcp/README.md Normal file
View File

@@ -0,0 +1,97 @@
# EVOLV MCP services — Docker stack
Three MCP services Claude Code uses to close real loops during EVOLV work:
| Service | What it does | Tools exposed |
|---|---|---|
| `mcp-node-red-admin` | Wraps the Node-RED HTTP admin API | `getFlows`, `postFlow`, `getFlow`, `inject`, `listNodes`, `restartFlow` |
| `mcp-influxdb` | Queries the telemetry bucket | `query`, `assertSeriesExists`, `assertRecentWrite`, `listMeasurements` |
| `mcp-browser` | Headless Playwright against the FlowFuse dashboard | `loadDashboard`, `screenshot`, `consoleLogs`, `waitForChart`, `getChartData` |
## Why these
Each closes a verification loop Claude currently cannot close:
- **Node-RED admin** — today Claude pushes flows via raw `curl`; the
MCP lets Claude deploy + fire injects + read live state in one
conversational turn.
- **InfluxDB** — today Claude cannot verify "did the telemetry land?"
beyond reading source. The MCP closes the loop after a deploy.
- **Browser** — today Claude cannot see the rendered dashboard. The
MCP catches the failure mode behind the η-null crash + the
blank-ui-chart bug + the editor pile-up bug at the only layer where
they're visible.
## Migration plan
These run locally **now** (we're in the middle of an infra migration).
Once the central MCP server is provisioned (target Q3 2026), each
service moves to shared infra by lifting the entry from this
`docker-compose.yml` plus the matching `tools/mcp/<name>/Dockerfile`
and pointing every developer's Claude Code at the central endpoint
instead of `localhost`. **The compose file stays here as the canonical
definition.**
## Usage
```bash
# build the three images (one-off, ~3 min)
cd tools
docker compose --profile mcp build
# start them
docker compose --profile mcp up -d
# wire Claude Code to them — add to your user-level .mcp.json
{
"mcpServers": {
"evolv-node-red-admin": { "type": "stdio", "command": "docker", "args": ["exec", "-i", "evolv-mcp-node-red-admin", "node", "server.mjs"] },
"evolv-influxdb": { "type": "stdio", "command": "docker", "args": ["exec", "-i", "evolv-mcp-influxdb", "node", "server.mjs"] },
"evolv-browser": { "type": "stdio", "command": "docker", "args": ["exec", "-i", "evolv-mcp-browser", "node", "server.mjs"] }
}
}
```
The repo-level `.mcp.json` is deliberately **not** committed (each
developer has different host endpoints / tokens). Use a user-level
config or `~/.claude.json`.
## Required environment
`tools/.env` (gitignored) with:
```dotenv
NODE_RED_HOST=http://host.docker.internal:1880
NODE_RED_TOKEN=… # optional, only if Node-RED has admin auth on
INFLUX_URL=http://host.docker.internal:8086
INFLUX_TOKEN=…
INFLUX_ORG=wbd
INFLUX_BUCKET=telemetry
DASHBOARD_URL=http://host.docker.internal:1880/dashboard
```
## Status
| Service | Dockerfile | Server impl | Status |
|---|---|---|---|
| `mcp-node-red-admin` | placeholder | **TODO** — see `mcp/node-red-admin/ROADMAP.md` | not runnable yet |
| `mcp-influxdb` | placeholder | **TODO** | not runnable yet |
| `mcp-browser` | placeholder | **TODO** — wrap `@playwright/test` | not runnable yet |
The compose file is the **target shape**. The Dockerfile + server
implementation per service is a follow-up (each is ~200400 LOC of MCP
protocol + the wrapped client). When a service lands, flip its row
above to `runnable` and remove the placeholder.
## When to use these — required reading
`CLAUDE.md` § "Tooling (Docker-first, local now, central later)" lists
the operating doctrine: **always prefer these tools over ad-hoc
curl/grep/manual checks**. Each tool exists because of a specific bug
class we've already paid for. Skipping them re-opens those bugs.
## Future: OPC-UA / PLC MCP
Out of scope for this round; will be revisited later. When added it
follows the same pattern: `tools/mcp/opcua/` with its own Dockerfile
and a row in this README.

View File

@@ -0,0 +1,5 @@
# placeholder — see ../README.md status table
FROM node:20-alpine
WORKDIR /app
RUN echo 'console.error("mcp-browser not yet implemented"); process.exit(1);' > server.mjs
CMD ["node", "server.mjs"]

View File

@@ -0,0 +1,5 @@
# placeholder — see ../README.md status table
FROM node:20-alpine
WORKDIR /app
RUN echo 'console.error("mcp-influxdb not yet implemented"); process.exit(1);' > server.mjs
CMD ["node", "server.mjs"]

View File

@@ -0,0 +1,5 @@
# placeholder — see ROADMAP.md
FROM node:20-alpine
WORKDIR /app
RUN echo 'console.error("mcp-node-red-admin not yet implemented — see ROADMAP.md"); process.exit(1);' > server.mjs
CMD ["node", "server.mjs"]

View File

@@ -0,0 +1,50 @@
# mcp-node-red-admin — implementation roadmap
Status: placeholder. Compose entry exists, server impl is TODO.
## Goal
An MCP stdio server that wraps the Node-RED admin HTTP API, exposing the
following tools to Claude Code:
| Tool | Wraps | Use case |
|---|---|---|
| `node_red_get_flows` | `GET /flows` | Read deployed flow JSON |
| `node_red_post_flow` | `POST /flow/:id` (single tab) | Deploy one tab without nuking others |
| `node_red_replace_flows` | `POST /flows` (bulk) | Replace the entire flow set |
| `node_red_inject` | `POST /inject/:nodeId` | Fire an inject node by id |
| `node_red_list_nodes` | `GET /nodes` | Discover registered node types |
| `node_red_restart_flow` | `DELETE` + redeploy | Force a restart of one tab |
## Sketch
```js
// tools/mcp/node-red-admin/server.mjs
import { Server } from '@modelcontextprotocol/sdk/server/index.js';
import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
const NODE_RED_HOST = process.env.NODE_RED_HOST;
const TOKEN = process.env.NODE_RED_TOKEN;
const server = new Server({ name: 'evolv-node-red-admin', version: '0.1.0' }, { capabilities: { tools: {} } });
server.setRequestHandler(/* listTools */);
server.setRequestHandler(/* callTool — fetch NODE_RED_HOST + tool's endpoint */);
await server.connect(new StdioServerTransport());
```
## Dockerfile sketch
```dockerfile
FROM node:20-alpine
WORKDIR /app
COPY package.json server.mjs ./
RUN npm install
CMD ["node", "server.mjs"]
```
## When to build it
After we've shipped enough EVOLV flow work that "deploy + fire inject +
read state in one turn" becomes the dominant inner loop. Today the
`curl` pattern still works for one-off deploys; the MCP earns its
keep when the loop runs 5+ times per session.

View File

@@ -0,0 +1,80 @@
# @evolv/physics-sanity
Cross-node physical-balance helpers. Import from any node's test files
to assert that scenario states close mass, hydraulic, hydraulic-power,
oxygen-transfer, or energy balances within a stated tolerance.
## Why
Per-node unit tests verify shape and behaviour. They don't catch
physically impossible plant states that arise from cross-node coupling
— e.g. a pumpingStation reporting outflow > inflow + accumulation, or a
diffuser reporting OTR inconsistent with its KLa × ΔC × V.
These helpers don't replace per-node tests. They sit on top of an
integration scenario and assert the closing balance.
## Usage
```js
const sanity = require('../../../tools/physics-sanity');
test('three-pump station closes the hydraulic balance', () => {
// … drive the scenario, take a snapshot …
const r = sanity.assertHydraulicBalance({
headerSuctionPa: ps.suctionPressurePa,
headerDischargePa: ps.dischargePressurePa,
pumpHeadPa: sumOfPumpHeads,
frictionPa: pipeFrictionEstimate,
});
assert.equal(r.ok, true, sanity.reportToString(r));
});
```
## Helpers exported
| Function | Asserts |
|---|---|
| `assertMassBalance({ inflowKgPerS, outflowKgPerS, accumulationKgPerS })` | `in - out - accumulation ≈ 0` |
| `assertHydraulicBalance({ headerSuctionPa, headerDischargePa, pumpHeadPa, frictionPa, staticHeadPa })` | `ΔP_headers ≈ pumpHead - friction - static` |
| `assertHydraulicPower({ flowM3PerS, headPa, shaftPowerW, efficiency })` | `shaft ≈ Q·H / η` |
| `assertOxygenTransfer({ klaPerS, csMgPerL, cMgPerL, otrKgPerS, volumeM3 })` | `OTR ≈ KLa · (Cs - C) · V` |
| `assertEnergyBalance({ heatInW, workInW, heatOutW, workOutW, accumulationW })` | `Q_in + W_in ≈ Q_out + W_out + ΔE` |
Each returns `{ ok, label, ...residuals }`. `reportToString(r)` formats
for human-readable failure messages.
## CLI demo
```bash
node tools/physics-sanity/bin/physics-sanity.js
```
Runs four sanity-check scenarios against the helpers (smoke-test for
the library itself).
## Tolerance defaults
| Domain | Absolute | Relative |
|---|---|---|
| mass | 1e-6 kg/s | 0.1 % |
| hydraulic ΔP | 50 Pa (0.5 mbar) | 0.1 % |
| hydraulic power | 1 W | 0.5 % |
| OTR | 1e-4 kg/s | 0.5 % |
| energy | 1 W | 0.1 % |
Override per call with `absTol` / `relTol`.
## Where to use this
Out-of-the-box destinations:
| Scenario | Where to add | Calls |
|---|---|---|
| pumpingStation hydraulic closure | `nodes/pumpingStation/test/integration/` | `assertHydraulicBalance`, `assertHydraulicPower` |
| reactor → settler mass balance | `nodes/reactor/test/integration/` | `assertMassBalance` |
| diffuser OTR vs reactor uptake | `nodes/diffuser/test/integration/` | `assertOxygenTransfer` |
| machineGroupControl efficiency sanity | `nodes/machineGroupControl/test/integration/` | `assertHydraulicPower` |
A future tool can scan integration tests and report which scenarios do
or don't have a closing-balance assertion.

View File

@@ -0,0 +1,18 @@
#!/usr/bin/env node
'use strict';
const lib = require('../index.js');
function runDemo() {
const checks = [
lib.assertMassBalance({ inflowKgPerS: 12.5, outflowKgPerS: 12.5, accumulationKgPerS: 0, label: 'reactor-passthrough' }),
lib.assertHydraulicBalance({ headerSuctionPa: 0, headerDischargePa: 110000, pumpHeadPa: 110000, label: 'station-A-headers' }),
lib.assertHydraulicPower({ flowM3PerS: 0.030, headPa: 110000, shaftPowerW: 5000, efficiency: 0.66, label: 'pump-A' }),
lib.assertOxygenTransfer({ klaPerS: 0.002, csMgPerL: 9.0, cMgPerL: 2.0, otrKgPerS: 2.8e-7, volumeM3: 20, label: 'diffuser-A' }),
];
for (const c of checks) process.stdout.write(lib.reportToString(c) + '\n');
const failed = checks.filter((c) => !c.ok).length;
process.exit(failed > 0 ? 1 : 0);
}
if (require.main === module) runDemo();

View File

@@ -0,0 +1,111 @@
'use strict';
const PA_PER_MBAR = 100;
const SECONDS_PER_HOUR = 3600;
function withinTolerance(observed, expected, absTol, relTol) {
if (!Number.isFinite(observed) || !Number.isFinite(expected)) return false;
const absErr = Math.abs(observed - expected);
if (absErr <= absTol) return true;
if (Math.abs(expected) > 0) return absErr / Math.abs(expected) <= relTol;
return false;
}
function assertMassBalance({ inflowKgPerS, outflowKgPerS, accumulationKgPerS = 0, label = 'mass', absTol = 1e-6, relTol = 1e-3 } = {}) {
const expected = inflowKgPerS - accumulationKgPerS;
const ok = withinTolerance(outflowKgPerS, expected, absTol, relTol);
return {
ok,
label,
inflowKgPerS,
outflowKgPerS,
accumulationKgPerS,
residualKgPerS: inflowKgPerS - outflowKgPerS - accumulationKgPerS,
relErr: expected === 0 ? null : (outflowKgPerS - expected) / expected,
};
}
function assertHydraulicBalance({ headerSuctionPa, headerDischargePa, pumpHeadPa, frictionPa = 0, staticHeadPa = 0, label = 'hydraulic', absTol = 50, relTol = 1e-3 } = {}) {
const lhs = headerDischargePa - headerSuctionPa;
const rhs = pumpHeadPa - frictionPa - staticHeadPa;
const ok = withinTolerance(lhs, rhs, absTol, relTol);
return {
ok,
label,
lhsPa: lhs,
rhsPa: rhs,
residualPa: lhs - rhs,
residualMbar: (lhs - rhs) / PA_PER_MBAR,
};
}
function assertHydraulicPower({ flowM3PerS, headPa, shaftPowerW, efficiency, label = 'hydraulic-power', absTol = 1, relTol = 5e-3 } = {}) {
if (!Number.isFinite(efficiency) || efficiency <= 0 || efficiency > 1.0) {
return { ok: false, label, msg: `efficiency=${efficiency} outside (0,1]` };
}
const expectedShaftPowerW = (flowM3PerS * headPa) / efficiency;
const ok = withinTolerance(shaftPowerW, expectedShaftPowerW, absTol, relTol);
return {
ok,
label,
flowM3PerS,
headPa,
efficiency,
expectedShaftPowerW,
observedShaftPowerW: shaftPowerW,
residualW: shaftPowerW - expectedShaftPowerW,
};
}
function assertEnergyBalance({ heatInW = 0, workInW = 0, heatOutW = 0, workOutW = 0, accumulationW = 0, label = 'energy', absTol = 1, relTol = 1e-3 } = {}) {
const inputs = heatInW + workInW;
const outputs = heatOutW + workOutW + accumulationW;
const ok = withinTolerance(inputs, outputs, absTol, relTol);
return {
ok,
label,
inputsW: inputs,
outputsW: outputs,
residualW: inputs - outputs,
};
}
function assertOxygenTransfer({ klaPerS, csMgPerL, cMgPerL, otrKgPerS, volumeM3, label = 'OTR', absTol = 1e-4, relTol = 5e-3 } = {}) {
if (!Number.isFinite(klaPerS) || klaPerS < 0) return { ok: false, label, msg: `KLa=${klaPerS} invalid` };
if (!Number.isFinite(volumeM3) || volumeM3 <= 0) return { ok: false, label, msg: `volume=${volumeM3} invalid` };
const driveMgPerL = csMgPerL - cMgPerL;
const expectedKgPerS = klaPerS * driveMgPerL * volumeM3 * 1e-3 / SECONDS_PER_HOUR * SECONDS_PER_HOUR / 1000;
const expectedKgPerS_corrected = klaPerS * driveMgPerL * volumeM3 / 1e6;
const ok = withinTolerance(otrKgPerS, expectedKgPerS_corrected, absTol, relTol);
return {
ok,
label,
klaPerS,
csMgPerL,
cMgPerL,
driveMgPerL,
volumeM3,
expectedKgPerS: expectedKgPerS_corrected,
observedKgPerS: otrKgPerS,
residualKgPerS: otrKgPerS - expectedKgPerS_corrected,
};
}
function reportToString(r) {
if (r.ok) return `OK ${r.label}`;
const fields = Object.entries(r)
.filter(([k]) => !['ok', 'label'].includes(k))
.map(([k, v]) => `${k}=${typeof v === 'number' ? v.toExponential(3) : v}`)
.join(' ');
return `FAIL ${r.label} ${fields}`;
}
module.exports = {
assertMassBalance,
assertHydraulicBalance,
assertHydraulicPower,
assertEnergyBalance,
assertOxygenTransfer,
reportToString,
PA_PER_MBAR,
};

View File

@@ -0,0 +1,14 @@
{
"name": "@evolv/physics-sanity",
"version": "0.1.0",
"private": true,
"description": "Cross-node physical-balance helpers (mass, hydraulic, energy). Import from test files; closure tolerance asserted at known plant states.",
"main": "index.js",
"bin": {
"evolv-physics-sanity": "bin/physics-sanity.js"
},
"scripts": {
"test": "node --test test/*.test.js"
},
"license": "UNLICENSED"
}

View File

@@ -0,0 +1,57 @@
'use strict';
const { test } = require('node:test');
const assert = require('node:assert/strict');
const lib = require('../index.js');
test('mass balance closes for steady-state pass-through', () => {
const r = lib.assertMassBalance({ inflowKgPerS: 10, outflowKgPerS: 10 });
assert.equal(r.ok, true);
assert.equal(r.residualKgPerS, 0);
});
test('mass balance reports residual when leaking', () => {
const r = lib.assertMassBalance({ inflowKgPerS: 10, outflowKgPerS: 9.5 });
assert.equal(r.ok, false);
assert.equal(Math.round(r.residualKgPerS * 1000), 500);
});
test('hydraulic balance: ΔP = pumpHead - friction - static', () => {
const r = lib.assertHydraulicBalance({
headerSuctionPa: 0,
headerDischargePa: 90000,
pumpHeadPa: 100000,
frictionPa: 8000,
staticHeadPa: 2000,
});
assert.equal(r.ok, true);
});
test('hydraulic power Q·H / η — within 0.5% relative tolerance', () => {
const Q = 0.03;
const H = 100000;
const eta = 0.65;
const shaft = (Q * H) / eta;
const r = lib.assertHydraulicPower({ flowM3PerS: Q, headPa: H, shaftPowerW: shaft, efficiency: eta });
assert.equal(r.ok, true);
});
test('hydraulic power flags eta=0', () => {
const r = lib.assertHydraulicPower({ flowM3PerS: 0.03, headPa: 100000, shaftPowerW: 5000, efficiency: 0 });
assert.equal(r.ok, false);
});
test('OTR check uses standard KLa formula', () => {
const kla = 0.002;
const cs = 9.0;
const c = 2.0;
const V = 20;
const otr = kla * (cs - c) * V / 1e6;
const r = lib.assertOxygenTransfer({ klaPerS: kla, csMgPerL: cs, cMgPerL: c, volumeM3: V, otrKgPerS: otr });
assert.equal(r.ok, true);
});
test('energy balance: heat-in + work-in = heat-out + work-out + accumulation', () => {
const r = lib.assertEnergyBalance({ heatInW: 1000, workInW: 200, heatOutW: 700, workOutW: 400, accumulationW: 100 });
assert.equal(r.ok, true);
});