pastilhasandClaude Fable 5
6071ded175
TODO.md had reached 1050 lines and 72 open items across 28 sections, read mostly on a phone. It is now a directory: todo/<theme>/<item>.md, one card per file, with frontmatter carrying id, status, created date and order, and the prose body copied across verbatim. TODO.md becomes a generated index (bun run todo:index). Dates: where an item's prose stated one it is used; otherwise its old section's; otherwise 2026-08-20, floored at the day before the split rather than invented. date_source records which. Shaped for a two-way Kanban board, filed as todo/tooling — a card is a file, id is the identity, frontmatter is the machine's half and the body is the owner's. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
1.9 KiB
id, title, theme, status, created, date_source, order
| id | title | theme | status | created | date_source | order |
|---|---|---|---|---|---|---|
| infra/nudger-a-control-in-the-chat-composer-beside | Nudger — a control in the chat composer, beside the background-job button | infra | open | 2026-08-20 | stated | 4 |
Nudger — a control in the chat composer, beside the background-job button. Wanted 2026-08-20 while leaving an agent to work overnight: "continue every hour if you have stalled".
scripts/nudge-agent.ts does it from outside tonight — it polls GET /api/chat/sessions?all=1
and speaks over the chat websocket, exactly as a browser does, so the platform needed no changes.
That is the workaround, not the feature.
Both things this said the platform lacked have since been built (verified 2026-08-21): the
recurring trigger is startNudger() at server.tsx:182, ticking once a minute over a nudges
table; the door into a session is startClaudeTurn({ ws: null }) at nudger.ts:102. What is left
is the UI — a control in the chat composer, beside the background-job button — plus generalising
delivery beyond the nudger's own schedule. See "Speaking into a session from the backend".
Design notes worth keeping:
- Never interrupt work. Check
isGeneratingimmediately before speaking. The point is to restart a stopped agent, not to hurry a busy one — the script already gets this right. - It must outlive what it nudges. A chat background job is owned by its session, so a nudge job launched from the session being nudged dies with it — the 30-minute idle GC, a clear, or a restart takes the watchdog with the agent. Whatever owns the schedule cannot be the session.
- Bound it. The script stops after
--max(12). A loop nobody is watching should end. - The same trigger would serve
COMMS/command-hq/watcher-loop-proposal.md, which wants infrastructure to own a loop rather than a model holding a protocol across turns.