Skip to content

Latest commit

 

History

History
88 lines (68 loc) · 3.57 KB

daily-workflow.md

File metadata and controls

88 lines (68 loc) · 3.57 KB

Daily workflow

Start here after the quickstart. These commands work in your brain repository with the installed brain CLI. If it is not on your PATH, use bun run brain instead. Capture and keyword search need no model account.

Capture and find a note

brain add "Odysseus needs timber and sailcloth for the raft." --title "Raft supplies"
brain search "sailcloth" --mode fts

On a fresh brain the capture creates notes/raft-supplies.md and indexes it. An occupied filename receives a suffix; use the path returned by your capture when reading or linking the note:

brain read notes/raft-supplies.md
brain add "Check [[notes/raft-supplies]] before leaving Ogygia." --title "Raft inspection"

The second note makes the connection explicit. Wiki-links resolve to files; concepts explains ambiguous names and directory links. You can keep everything as note until a useful taxonomy emerges. /brain-init or hand-edited configuration can add types later. /process-notes helps an agent review and file the inbox.

Review changes and index edits

brain add and MCP writes update the index themselves. After editing markdown in an editor, refresh it explicitly:

brain index
brain validate
brain audit
brain stats

Validation checks file/config correctness. Audit reports maintenance findings; its informational findings are distinct from errors and warnings. Stats shows current health, the thresholds in force and inventory. Read the findings before changing content. After reviewing a mechanical edit that deliberately did not change updated, brain accept-mtime accepts the new file timestamp as the baseline for silent-edit detection.

To retire the inspection note without deleting its history:

brain archive notes/raft-inspection.md

Archived content is excluded from default search; brain search "Ogygia" --include-archived includes it. Use the actual capture path if a suffix was added.

Maintain and sync

brain maintain --no-git runs the content/index maintenance steps without git packing. It can regenerate declared registries, refresh the index and prune disposable scratch files; inspect its report for failed steps. The CLI reference describes the sequence and flags.

A private remote gives you an off-machine copy. Configure and verify it before syncing; the backup guide explains what needs to be kept. Run sync from the main branch of a brain whose origin is configured:

brain sync --human

Sync commits recognized content, pulls, merges and pushes. It leaves unknown files and unresolved conflicts for judgment. It can invoke your configured agent runner when that judgment is needed; without the required runner/account, complete the handoff rather than treating an unfinished sync as success. The sync reference explains machine output and exit codes, and configuration explains pull and merge policy.

Choose the next capability

  • Modules add a domain workflow when your notes need one.
  • MCP gives an agent structured access to the same files and search.
  • Semantic search adds vectors when you choose to provide credentials; keyword search remains available.
  • Search evaluation measures whether retrieval changes improve answers on your own query set.
  • Hosting keeps backup and optional phone access separate from the local workflow.