Files
app/NOTEPAD.md
T
stroblmeandClaude Fable 5 be6d4fc13c Build dashboards you can actually look at and press
Widgets bind to a message name and read it live off the socket the editor
already had — lifted out of the flow editor so a dashboard route gets the
same values, which also gives the home page live data for free.

The input widgets close the loop the other way: a slider publishes into
the graph and whatever consumes that message runs. Verified end to end in
the running app — moving a slider set a flow input, and the stat bound to
what the flow computed from it followed.

View mode is plain CSS grid. A wall panel that only displays should not
download the code that lets someone drag things around, and it now does
not. Editing is a widget picker, a per-widget width control and a
settings card fed by the message catalog.

No new dependencies: the slider is a range input, the gauge is an arc,
and the markdown is a five-line subset. Charts are the one widget still
missing — they need a charting library and the chart tokens the design
guidelines reserved — so they are stored and validated but not offered.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011LF61rxW1FG5YCD2J9YqjY
2026-08-16 09:33:02 +02:00

7.6 KiB

This file captures tasks which derive from roadmap tasks (unfinished, deferred), bugs encountered during usage and feature requests/improvements which are not fitting directly in the roadmap. Always sort by priority and put tasks blocked by other tasks/features at the dedicated section. When working on a task, check for other, similar tasks that could be resolved on the way. Use following pattern to classify tasks: TYPE/SCOPE Where TYPE could be BUG, FEAT, PERF, CHORE and SCOPE could be UX, UI, FLOW, NODE, API, INFRA, DOCS appended by MOBILE if only for mobile use case. Don't write temporary reasons for deferring a task in the task description (only strategical reasons should be noted). Deferring because out of scope is fine, but don't mention deferring than.

Open

  • BUG/UI: when the edges pulse on message, they change their color (go from slightly darker back to normal) at the end again. Double check that after the animation they end up with the same color as before to not have this double-pulse effect

  • BUG/UI: enlarge the icon in the sidebar slightly

  • BUG/UI: clicking outside the panel does not discard the flow edit panel

  • BUG/UI: the graph showed in the node edit panel should also be shown for a specific edge inside the pop-up panel when clicking the edge

  • FEAT/FLOW: single-stepping a paused flow. The work queue it needed now exists: pausing parks claimed items per flow and resuming drains them, so a step button is a matter of popping one parked item instead of all of them.

  • FEAT/UI: interrupting a node that is already running. Pause holds nodes that have not been submitted yet; one already executing runs to completion.

  • BUG/UI: the enlarged panel (for code editing) should still maintain its floating style

  • BUG/UI: SidePanel's mobile branch does not set data-testid on the SheetContent, so [data-testid=node-panel] does not exist on a phone. Mobile specs cannot address the panel.

  • CHORE/API: node source saves carry no version precondition, so two clients editing the same node's code are last-writer-wins. The flow document is what the optimistic lock protects; code files would need their own, and an exact-match one produces false conflicts against a single client's own interleaved flow and source saves.

  • FEAT/UI: publishing and discarding are only reachable while no side panel is open, since the floating chrome hides for the panel. Editing a node's code and publishing it is therefore close-panel-then-publish.

  • FEAT/UI: no screen for the OAuth clients an agent registers. They can only be listed or removed in the database, so withdrawing one agent's access means deleting its rows or rotating OAUTH_PRIVATE_KEY_FILE, which cuts off all of them.

  • CHORE/INFRA: requires-python is capped below 3.14 because the MCP SDK wants a newer starlette there than the pinned sentry-sdk<2 allows. Lift the cap when sentry-sdk moves to 2.x.

  • FEAT/UI: no screen for alerting. Channels and rules are API-only (/alerts/config), so setting up a phone notification means calling the endpoint by hand.

  • FEAT/UI: there is no screen for managing the secrets store itself. A node parameter marked x-secret offers the stored secrets, but they can only be created through the API.

  • FEAT/FLOW: input discretization drops the trailing edge — if a producer goes quiet inside the window, the consumer never sees the last value. Delivering it at the end of the window needs a timer per (node, port).

  • CHORE/FLOW: shared node sources bypass the draft/publish split. Editing one writes the library copy and reloads immediately, since the code is not any single flow's to hold back. Deliberate, but it means a shared node is the one thing publish does not gate.

  • FEAT/UI: flow-level edits (title, inputs) go through FlowPanel.onChange and stay outside the undo stack.

  • CHORE/UI: make test-backend cannot reach Postgres while the integrated stack is up — compose.local.yml does db: ports: !reset []. Run it against the container's address, or move the suite inside the compose network.

  • CHORE/UI: the Playwright specs run against the development stack and leave their users and flows behind, which is why the flowbar filled with test_flow_*. Give them their own data or clean up after themselves, as pytest now does.

  • CHORE/UI: the standalone Vite origin (make dev-frontend on :5173) is still not in BACKEND_CORS_ORIGINS. scripts/setup.sh regenerates the list on every make init and compose.local.yml re-declares it as a container variable that wins, so the origin has to be added in both places to work.

  • CHORE/INFRA: test-backend.yml runs pytest on the runner host, so it keeps the fixed container name and port 5432 that compose.ci.yml resets everywhere else. Two concurrent runs of that one job collide.

  • PERF/UI: (deferred for now) the Monaco chunk is 2.6 MB. It only loads when a node panel opens, but the editor could be trimmed further or swapped for CodeMirror if that becomes a problem.

  • CHORE/FLOW: four modules of app/flow/nodes/ are still excluded from strict mypy ([[tool.mypy.overrides]] in pyproject.toml): mqtt, http, influx and delay. They carry prototype typing, croniter ships no stubs and influxdb_client does not re-export its names. base and mlp are checked; shrink the rest as each integration is revisited.

  • CHORE/FLOW: a node function returning something other than a dict raises AttributeError in Node._to_messages rather than a named error. Outputs are keyed by port, so a non-dict cannot be one — say so where the return value is mapped.

  • PERF/FLOW: every save rebuilds the whole pipeline. Fine at the current flow count; rebuild only the touched flow when it starts to show.

  • FEAT/UI: the chart widget is stored and validated but not drawn — it is filtered out of the editor's add row. Needs a charting library and the --chart-* tokens below.

  • FEAT/UI: dashboard widgets are sized with a wider/narrower control rather than dragged. A grid library would give drag-and-resize and per-breakpoint layouts; the document already stores layout per breakpoint, so only the editor changes.

  • FEAT/UI/MOBILE: dashboards render in the padded admin shell, so a wall panel gets the sidebar, footer and a max-w-7xl column. A full-bleed shell would suit a panel better.

  • FEAT/UI: reintroduce --chart-* tokens as one designed sequential scale when the first chart lands. The node sparkline draws one series in --primary and needs none.

  • PERF/UI: the app's entry chunk is 680 kB (210 kB gzipped) and exceeds the warning threshold. React Flow and Monaco are already lazy; a manualChunks split measured no better, so this needs route-level work on the shell rather than chunking config.

  • CHORE/INFRA: bun run --filter frontend build fails on this workspace with crypto.hash is not a function — Vite 7 wants Node 20.12+ and the host has 18. The Docker image builds fine, so it only bites local bundling; bunx tsc still type-checks.

Blocked

  • FEAT/UI: a "Bug" icon on the node error bubble opening the logs panel at that node's traceback. The panel now has the traceback; wiring the bubble to open and filter it needs the panel's open state lifted into FlowEditor.
  • FEAT/INFRA: MQTT broker and InfluxDB compose services for local development. The node types exist; a local broker would make them testable without external hardware.
  • CHORE/INFRA: bun install inside the frontend Docker build intermittently fails with "Fail extracting tarball" for several packages at once, and succeeds on a plain rebuild. It looks like concurrent extraction under memory pressure. Pin down or retry in the Dockerfile if it starts costing CI time.