Skip to content

docs: scheduling + digest catch-up (three platforms, delivery layers, Telegram) #239

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
kengio merged 1 commit into from
Jul 30, 2026
Merged

docs: scheduling + digest catch-up (three platforms, delivery layers, Telegram) #239

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@
</p>

<p align="center">
<strong>OneBrain</strong> is a free, open-source AI OS layer: persistent memory, 31 skills, and a portable vault that works with any AI harness — entirely on your machine.
<strong>OneBrain</strong> is a free, open-source AI OS layer: persistent memory, 31 skills, scheduled skills that report to your vault and Telegram, and a portable vault that works with any AI harness — entirely on your machine.
</p>

<p align="center">
Expand Down Expand Up @@ -218,9 +218,11 @@ One driver profile, every garage. OneBrain Sync will keep your vault, memory, an

## Scheduling

Run OneBrain skills automatically — daily briefings, weekly reviews, recurring maintenance — via your OS scheduler (macOS launchd; Linux + Windows coming soon), configured in `onebrain.yml`.
Run OneBrain skills automatically — a morning digest (`/digest`: markets, news, Reddit — you pick the sections in a one-time interview), daily briefings, weekly reviews, recurring maintenance — via your OS scheduler on **macOS (launchd), Windows (Task Scheduler), and Linux (systemd user timers)**, configured in `onebrain.yml`.

Full config format, preset bundles, and CLI flags → [docs/scheduling.md](docs/scheduling.md)
Every scheduled run writes its output to your vault's skill log, and — if you set `notifications.telegram_chat_id` — also sends it to your Telegram, so the morning digest and daily briefing land on your phone before you open a laptop.

Full config format, delivery layers, preset bundles, and CLI flags → [docs/scheduling.md](docs/scheduling.md)

## Customization

Expand Down
24 changes: 19 additions & 5 deletions docs/scheduling.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,10 +4,12 @@ Run OneBrain skills automatically on a recurring or one-shot schedule via your O

> Part of [OneBrain docs](README.md)

OneBrain skills can run automatically on a schedule via your OS scheduler (macOS launchd; Linux + Windows coming soon). Configure in `onebrain.yml`:
OneBrain skills can run automatically on a schedule via your OS scheduler — **all three platforms since CLI v3.4.20**: macOS (launchd), Windows (Task Scheduler), Linux (systemd user timers). Every backend registers into your *user* session and is fire-verified on real machines; per-platform semantics (artifacts, missed-run behavior, output capture) are documented in the CLI's [platform-support](https://github.com/onebrain-ai/onebrain-cli/blob/main/docs/platform-support.md). Configure in `onebrain.yml`:

```yaml
schedule:
- cron: "30 8 * * *" # morning digest 8:30 (markets/news/Reddit — see /digest)
skill: /digest
- cron: "0 9 * * *" # daily 9am
skill: /daily
harness: codex
Expand All @@ -29,7 +31,7 @@ schedule:
skill: /daily
```

After firing, the launchd plist auto-uninstalls itself.
After firing, the scheduler artifact deletes itself — on every platform (launchd self-removal · Task Scheduler `DeleteExpiredTaskAfter` · systemd `ExecStopPost`).

Register schedules:

Expand All @@ -46,7 +48,19 @@ Or use the interactive wizards from inside your vault:
/schedule-remove # remove an entry
```

Output goes to `[logs_folder]/scheduler/YYYY/MM/YYYY-MM-DD-{skill}.md` as readable markdown.
## Where scheduled output goes

Three layers (CLI v3.4.20 + plugin v3.4.7):

1. **Vault skill log** — the run's primary output is appended to `[logs_folder]/log/YYYY/MM/YYYY-MM-DD-{skill}.md`: searchable, `/recap`-visible, readable anywhere your vault syncs. This is the deliverable.
2. **Telegram (opt-in)** — set `notifications.telegram_chat_id` in `onebrain.yml` and every scheduled skill run also sends its output to that chat. Best-effort: a send failure never fails the run. Setup is a capture, not a hunt — `/digest` (first run), `/schedule-add`, and `/onboarding` all offer it when the Telegram channel is configured: you send the bot one message and OneBrain records the chat id for you.

```yaml
notifications:
telegram_chat_id: "123456789" # unset = vault log only
```

3. **Raw process logs** — stdout/stderr land in the OS state dir (`~/Library/Logs/onebrain` on macOS · journald on Linux · Task Scheduler history on Windows), deliberately **outside** the vault: a cloud-synced vault once made launchd fail every run silently (onebrain-cli #315). Failures also write `[logs_folder]/scheduler/**/*.err.md`, which `/doctor` and session startup surface.

## Command mode (CLI binaries, hook-style)

Expand Down Expand Up @@ -78,9 +92,9 @@ CLI flags:

| Flag | Purpose |
|---|---|
| `--dry-run` | Print plist without writing |
| `--dry-run` | Print the scheduler artifact(s) without writing |
| `--remove` | Remove all OneBrain schedules |
| `--refresh` | Re-emit plists after vault move |
| `--refresh` | Re-emit scheduler artifacts after a vault move |
| `--resume <skill>` | Resume an auto-paused skill |
| `--status` | Show registered schedules + run history |
| `--test <skill>` | Manually invoke a scheduled skill once |
Expand Down
Loading