Front-loads gap discovery before /grill-me by adding two skills:
research MOSTLY fans out Explore + WebSearch agents in parallel,
synthesizes findings into a brief, names open
unknowns explicitly (which become /prototype targets)
prototype MOSTLY builds a throwaway spike to test ONE falsifiable
assumption; code lives in .prototypes/ (gitignored),
never promoted; output is evidence — verdict, numbers,
observed behavior — that feeds /prd
Full chain now:
/research → /prototype → /grill-me → /prd → /prd-to-issues → /ship-it
Chain rationale: /research and /prototype surface knowledge gaps and falsify
risky assumptions while the cost of changing direction is still cheap; the
TOGETHER phases (grill-me, prd) lock down the contract; the AFK phase
(ship-it) only executes against contracts already on paper.
The chain is a default, not a mandate — README covers when to skip
upstream skills for small or stack-familiar work.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Four workflow skills that take a feature from fuzzy idea to merged code.
Two human-in-the-loop phases (grill-me, prd), one mostly-together (prd-to-issues
files only on explicit 'create'), and one AFK (ship-it).
grill-me TOGETHER pressure-test the idea with hard interview questions
prd TOGETHER synthesize PRD; gaps stay explicit, not papered over
prd-to-issues MOSTLY thin vertical-slice issues with coverage matrix +
per-issue Slice check; self-audits before showing
ship-it AFK shell loop ships each slice end-to-end with one
commit per issue, status streams to terminal,
Ctrl-C-able, survives session close
Vertical-slice principle throughout: every issue cuts end-to-end through every
integration layer (no horizontal "do all the DB work first" issues). The
AFK loop only ships against acceptance criteria already locked in by the PRD
phase — autonomous code never runs against undefined contracts.
ship-it tracker support: gh (GitHub) and tea (Gitea). For this repo, set
SHIP_IT_TRUNK=development to override the main default.
See .claude/skills/README.md for the full how-to and a worked example.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PALETTE REDESIGN (2026-05-21)
Sidebar swatches switched from S88 level (all blue) to domain-hue per node.
Family hue = function (rotating=orange, valves=teal, biology=green/olive,
sampling=violet, sensor=amber, aeration=sky-blue, infrastructure=slate);
within a family, darker = higher S88 / "more controller-ish."
Editor-group rectangles in flow.json still follow S88 — only the
registerType colour changed.
Submodule bumps for palette: rotatingMachine, machineGroupControl,
pumpingStation, valve, valveGroupControl, reactor, settler, monster,
measurement, diffuser, dashboardAPI.
Docs touched:
- CLAUDE.md: palette swatch vs. editor-group bullets split out.
- .claude/rules/node-red-flow-layout.md: new §10.0 introduces the two
color systems, full 12-row palette table, and explicit warning not to
mix the two hexes.
- .claude/refactor/MODULE_SPLIT.md: per-node headers annotated with
both `group #XXX` and `palette #XXX`.
- .claude/refactor/WIKI_HOME_TEMPLATE.md + WIKI_TEMPLATE.md: clarify
Mermaid classDefs visualize hierarchy, not palette swatches.
- .claude/refactor/OPEN_QUESTIONS.md: dated decision entry with
rationale, file list, and follow-ups.
CORESYNC SUBMODULE (new)
nodes/coresync added pointing at https://gitea.wbd-rd.nl/RnD/coresync.
FROST/SensorThings handoff path — first version forwards FROST-ready HTTP
request messages on the dbase output; a downstream http-request node
performs the POST and feeds responses back on msg.topic = "frost.response".
Lazy stream resolver, latest-wins queue (keep first + latest, drop middle),
knot-emit on slope change, provenance preserved in Observation parameters.
- .gitmodules: add nodes/coresync entry.
- package.json: register coresync as a Node-RED node.
- generalFunctions bump: new frostFormatter + 4 node config schemas
expose the dbase format option.
- measurement bump: "frost" option added to dbaseOutputFormat dropdown
(plus the in-flight data.measurement unit-handling work).
- machineGroupControl bump: small editor compact-fields tweak alongside
the palette change.
- CORESYNC_FROST_INTERVIEW_HANDOFF.md added at root with interview state
(Q20 open: slope angle vs. relative delta comparison).
DASHBOARDAPI MODULE_NOT_FOUND FIX
package.json: dashboardapi entry path corrected to
nodes/dashboardAPI/dashboardAPI.js. Commit e04c4a1 renamed the files to
camelCase but missed package.json; on case-sensitive filesystems
(Linux/Docker, where the tarball lands) the require resolved to nothing
and the node showed MODULE_NOT_FOUND in the Node-RED palette.
MISC CLEANUP
- examples/README.md + examples/pumpingstation-complete-example/ removal
(build_flow.py, flow.json, README.md superseded by per-node examples).
- jest.config.js: in-progress tweak.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Delete .mcp.json + .claude/rules/repo-mem.md; drop .repo-mem from .gitignore
- Remove repo-mem / substrate_score / repo_search references from all .md
- Move 15 EVOLV skills from .agents/skills/ to .claude/skills/ so they are
auto-discovered by the Claude Code harness and invokable via the Skill tool
- Retire .agents/skills/evolv-orchestrator (duplicate of the subagent at
.claude/agents/evolv-orchestrator.md); orchestrator lives as a subagent only
- Drop OpenAI-format agent yaml metadata from each skill (not needed for CC)
- Update CLAUDE.md, CONTRACTS.md, AGENTS.md to point at the new locations and
disambiguate skills (.claude/skills/) vs subagents (.claude/agents/)
- Fix CLAUDE.md tick-loop wording (opt-in per-node, not a fixed 1000ms)
- Widen .claude/rules/ paths frontmatter so node-architecture and telemetry
rules trigger on more relevant files; add frontmatter to flow-layout rule
- Bump CONTRACTS.md review date to 2026-05-19; add step 7 to the contract-
change workflow (review example flows when topic usage changes)
- Bump nodes/generalFunctions pin (Home.md substrate_score reference removed)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Establish CONTRACTS.md at the EVOLV root as the canonical map of where every
contract, rule, and standard lives. Surface it from CLAUDE.md so every fresh
agent or colleague lands there first.
Reshape .claude/refactor/ to reflect that the platform refactor is done:
live standards stay at the top level; the plan artifacts (CONTINUE_HERE.md,
TASKS.md) move into Archive/ with WARNING banners.
Drop content that drifted out of date or duplicated the new standards stack:
- docs/DEVELOPER_GUIDE.md (pre-refactor walkthrough; superseded by
wiki/Architecture, wiki/Getting-Started, .claude/rules/node-architecture,
.claude/refactor/MODULE_SPLIT + per-node CONTRACT.md + src/commands/).
- .agents/decisions/ (15 DECISION files): load-bearing decisions belong in
commit messages and PR descriptions; live open items in OPEN_QUESTIONS.md.
- .agents/improvements/TOP10_*.md: moved to Archive/.
Bump generalFunctions to 49c77f2 — adds CONTRACT.md inside the library:
different shape from per-node CONTRACT.md files (library API, not msg.topic),
with stability tags and pointers to .claude/refactor/CONTRACTS.md §N.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
MGC submodule lands the 2026-05-14 governance review fixes plus rolled-up
session work: _output-manifest.md per the new output-coverage rule,
computeEqualFlowDistribution extracted as a pure function (testable without
MGC), groupEfficiency degenerate-case fix, unit-self-describing set.demand,
eta = (Q*dP)/P formula correction, and dashboard fan-out hardening
(auto-init, NCog normalization, Q-H trim, null-trap closure). Suite 108/108.
Superproject adds:
- .claude/rules/output-coverage.md: every-output-every-state testing rule
prompted by the eta-null crash earlier in the session.
- CLAUDE.md: pointer to the new rule under Conventions.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
CONTRACTS.md §4: full payloadSchema.type table including 'none', plus
the optional description field example. Matches the B3.2 implementation.
WIKI_TEMPLATE.md §5: Unit column appears with explanatory paragraph.
Matches the P11.4 wikiGen output.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Single doc capturing the 'what's not done' at the end of the
2026-05-11 sprint, in priority order: B5 reactor boundary-conditions
merge, Phase 8 PR cycle, a handful of small open-questions follow-ups,
plus the wiki cosmetics list. README.md links to it from the top.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
All 11 nodes' wiki/Home.md regenerated with the Unit column +
per-topic descriptions. rotatingMachine + reactor private-method
test files rewritten to the public BaseNodeAdapter surface.
OPEN_QUESTIONS: rotatingMachine + reactor private-test entries
marked RESOLVED.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
WIKI_TEMPLATE.md — extend the canonical per-node page from 9 to 14 sections:
+ Header band (commit hash + regen date)
+ Code map (flowchart TB w/ subgraphs over concern modules)
+ Child registration (mirrors ChildRouter declarations)
+ Data model — getOutput() (abstract schema + optional concrete sample)
+ Debug recipes (symptom → first thing to check)
+ AUTOGEN markers around topic-contract + data-model schema so the
Phase 9 regen script can rewrite in place.
+ 'Picking a visual' table: Mermaid is default, plots/SVG/screenshots
allowed where they serve the data.
+ Archive banner snippet.
WIKI_HOME_TEMPLATE.md (new) — Home.md + Archive.md templates:
- Platform-wide Mermaid graph of 11 active nodes, S88-coloured.
- Navigation table grouped by S88 level.
- Standards-pointer table to .claude/rules + .claude/refactor docs.
- Live refactor-status table for returning visitors.
- Archive index template with archival-date column.
No wiki pages written yet — next step is one worked example
(pumpingStation) before any change to the Gitea wiki repo.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Defines the 9-section template every node's wiki page follows:
1. Position in the platform (mermaid flowchart, S88-coloured)
2. Capability matrix (≤ 10 rows)
3. Topic contract (auto-generated from src/commands/index.js)
4. Lifecycle (mermaid sequenceDiagram)
5. Configuration (mermaid flowchart + form-to-config table)
6. Examples (basic/integration/dashboard tiers)
7. State chart (stateDiagram-v2, only for stateful nodes)
8. When you would NOT use this node
9. Known limitations / current issues
Hard rules: diagrams before prose, ≤ 60 words of unbroken prose
anywhere, topic contract auto-generated (no hand-written drift).
Per-node application is the next step (P9.3-P9.6 in TASKS.md).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Resolves the 5 open questions answered during Phase 1 setup:
- Topic naming: canonical from Phase 1 (set/cmd/data/child/query/evt),
with full glossary in CONTRACTS.md §1.
- Parent EVOLV branch lineage: rebased onto origin/main.
- Deprecated paths: tracked as Phase 8.5 in TASKS.md.
- Child storage: registry-as-truth + named getters via
declareChildGetter.
- Tick: opt-in via static tickInterval; default is event-driven via
source.emitter 'output-changed'. statusInterval (always-on, 1Hz)
is separate.
Plus two new pre-existing-issue notes from the sanity gate:
- dashboardAPI uses Mocha-style describe() under node:test (broken).
- reactor tests are mathjs-bound (~13s/file load).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Platform-wide refactor plan: README, CONVENTIONS, CONTRACTS,
MODULE_SPLIT, TASKS, OPEN_QUESTIONS. Source of truth for the
phased refactor across all 12 submodules.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- package.json: remove @tensorflow/tfjs and @tensorflow/tfjs-node.
Monster's TF code was already stripped; the deps were stale and kept
pulling a heavy native binary back into every install.
- .gitignore: ignore .repo-mem/ regenerable indexes and per-session
.claude/*.lock runtime files.
- CLAUDE.md: prepend READ-FIRST pointer to .claude/rules/repo-mem.md;
collapse the 'three outputs' bullet to a pointer at node-architecture.
- .claude/rules/telemetry.md: drop Port 0/1/2 duplication; reference
node-architecture.md.
- .claude/rules/testing.md: stop requiring a separate test/edge tier and
the basic/integration/edge example flow trio. Reflects what nodes
actually do.
- .claude/rules/repo-mem.md (new): when-to-call-which guide for the
per-repo memory MCP, anti-patterns, refresh model.
- .mcp.json (new): wire repo-mem stdio server.
- docs/DEVELOPER_GUIDE.md (new): step-by-step guide for adding a new
EVOLV node under the three-layer pattern.
- Bump nodes/pumpingStation to 6ab585b (docs + simulations refresh,
spill-flow path renames consistent with d8490aa).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Basin undersized (10m³) for sinus peak (126 m³/h) → overflow → 122%.
Now 30 m³ with 4m height, all PS fields set. New rule: always configure
every field of every node.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Charts rendered blank because the helper was missing 15+ required
FlowFuse properties. The critical three:
- interpolation: "linear" (no line drawn without it)
- yAxisProperty: "payload" + yAxisPropertyType: "msg" (chart didn't
know which msg field to plot)
- xAxisPropertyType: "timestamp" (chart didn't know the x source)
Also: width/height must be numbers not strings, colors/textColor/
gridColor arrays must be present, and stackSeries/bins/xAxisFormat/
xAxisFormatType all need explicit values.
Fixed the ui_chart helper to include every property from the working
rotatingMachine/examples/03-Dashboard.json charts. Added the full
required-property template + gotcha list to the flow-layout rule set
(Section 4) so this class of bug is caught by reference on the next
chart build.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sections 10-16 extend the existing flow-layout rule with a deterministic
lane-and-group convention anchored in the S88 hierarchy:
- 8 logical lanes: L0 inputs -> L1 adapters -> L2 CM -> L3 EM -> L4 UN
-> L5 PC -> L6 formatters -> L7 outputs. 240 px between lanes.
- Lane assignment is by S88 level, not by node name. New nodes inherit
a lane via a NODE_LEVEL registry, no rule change needed.
- Every parent + its direct children is wrapped in a Node-RED group box
coloured by the parent's S88 level (Pump A = EM blue, MGC = Unit blue,
PS = Process Cell blue, ...). Search the parent's name -> group
highlights.
- Utility clusters (mode broadcast, station-wide commands, demand
fan-out) use neutral-grey group boxes.
- Dashboard / setup / demo-driver tabs each get a variant of the rule.
- Spacing constants, place() and wrap_in_group() helpers, an 8-step
verification checklist.
Off-spec colours (settler orange, monster teal, diffuser and
dashboardAPI missing) are flagged in Section 16 as a follow-up cleanup.
The NODE_LEVEL registry already maps those nodes to their semantic S88
level regardless of what the node's own colour currently says.
Rule lives in the superproject only; per-node repos will reference it
from their own CLAUDE.md files (separate commits per submodule).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The demo was a single 96-node tab with everything wired directly. Now
4 tabs wired only through named link-out / link-in pairs, and a
permanent rule set for future Claude sessions to follow.
Tabs (by concern, not by data flow):
🏭 Process Plant only EVOLV nodes (3 pumps + MGC + PS + 6 measurements)
+ per-node output formatters
📊 Dashboard UI only ui-* widgets, button/setpoint wrappers, trend
splitters
🎛️ Demo Drivers random demand generator + state holder. Removable
in production
⚙️ Setup & Init one-shot deploy-time injects (mode, scaling,
auto-startup, random-on)
Cross-tab wiring uses a fixed named-channel contract (cmd:demand,
cmd:mode, cmd:setpoint-A, evt:pump-A, etc.) — multiple emitters can
target a single link-in for fan-in, e.g. both the slider and the random
generator feed cmd:demand.
Bug fixes folded in:
1. Trend chart was empty / scrambled. Root cause: the trend-feeder
function had ONE output that wired to BOTH flow and power charts,
so each chart received both flow and power msgs and the legend
garbled. Now: 2 outputs (flow → flow chart, power → power chart),
one msg per output.
2. Every ui-text and ui-chart fell on the (0, 0) corner of the editor
canvas. Root cause: the helper functions accepted x/y parameters
but never assigned them on the returned node dict — Node-RED
defaulted every widget to (0, 0) and they piled on top of each
other. The dashboard render was unaffected (it lays out by group/
order), but the editor was unreadable. Fixed both helpers and added
a verification step ("no node should be at (0, 0)") to the rule set.
Spacing convention (now codified):
- 6 lanes per tab at x = [120, 380, 640, 900, 1160, 1420]
- 80 px standard row pitch, 30-40 px for tight ui-text stacks
- 200 px gap between sections, with a comment header per section
New rule set: .claude/rules/node-red-flow-layout.md
- Tab boundaries by concern
- Link-channel naming convention (cmd:/evt:/setup: prefixes)
- Spacing constants
- Trend-split chart pattern
- Inject node payload typing pitfall (per-prop v/vt)
- Dashboard widget rules (every ui-* needs x/y!)
- Do/don't checklist
- Link-out/link-in JSON cheat sheet
- 5-step layout verification before declaring a flow done
CLAUDE.md updated to point at the new rule set.
Verified end-to-end on Dockerized Node-RED 2026-04-13: 168 nodes across
4 tabs, all wired via 22 link-out / 19 link-in pairs, no nodes at
(0, 0), pumps reach operational ~5 s after deploy, MGC distributes
random demand, trends populate per pump.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Update all submodule URLs from gitea.centraal.wbd-rd.nl to gitea.wbd-rd.nl
- Add settler as proper submodule in .gitmodules
- Add agent skills, function anchors, decisions, and improvements
- Add Docker configuration and scripts
- Add manuals and third_party docs
- Update .gitignore with secrets and build artifacts
- Remove stale .tgz build artifact
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
