Skip to content

A self-organizing Obsidian vault that gives AI coding agents persistent memory.

12.1 KB
Raw Blame

date: 2026-04-07 description: "Vault-specific workflows and slash commands — reusable patterns for review prep, project tracking, and vault maintenance" tags: - brain - index

Skills

Custom slash commands, subagents, and reusable workflows. Defined in .claude/commands/ and .claude/agents/. This note is the canonical command catalog for agents that don't get the in-session skills injection (Codex, Gemini, Cursor) — keep it current when commands change; AGENTS.md deliberately carries no command table.

Slash Commands

Daily Workflow

Command Purpose
/om-standup Morning kickoff — load context, review yesterday, surface tasks, identify priorities
/om-dump Freeform capture — dump anything, gets routed to the right notes automatically
/om-wrap-up Full session review — verify notes, indexes, links, suggest improvements. Auto-triggered on "wrap up".

Editing & Synthesis

Command Purpose
/om-humanize Voice-calibrated editing — makes Claude-drafted text sound like you wrote it
/om-weekly Weekly synthesis — cross-session patterns, North Star alignment, uncaptured wins

Meeting Prep & Capture

Command Purpose
/om-prep-1on1 Prep for an upcoming 1:1 — load person context, open items, wins to share, suggested agenda
/om-capture-1on1 Capture 1:1 meeting transcript into structured vault note with quotes, action items, DM context
/om-incident-capture Capture incident from Slack channels/DMs into structured vault notes — timeline, people, analysis, brag entry
/om-slack-scan Deep scan Slack channels/DMs for evidence — extracts timestamped touchpoints, organizes by context
/om-meeting Prep for any meeting by topic — subject-forward briefing with open items, blockers, and brainstormed considerations
/om-intake Process meeting notes inbox — classify raw exports in work/meetings/ and route to the right vault notes

Performance & Review

Command Purpose
/om-peer-scan Deep scan a peer's GitHub PRs for review prep — produces structured analysis saved to perf/evidence/
/om-review-brief Generate review brief (manager or peer version) from vault data
/om-self-review Write self-assessment for review tool — projects, competencies, principles
/om-review-peer Write peer review — projects, principles, performance summary

Vault Maintenance

Command Purpose
/om-correct Sweep a corrected fact through the vault — semantic + literal search, applies at the single source, replaces restatements with links, preserves notes recording what was believed at the time. The acting half of Write-Correctness Law 2
/om-vault-audit Deep structural audit — indexes, frontmatter, links, Bases, folder placement, stale context
/om-vault-upgrade Upgrade this vault through ShardMind, keeping local improvements and listing them as upstream candidates; or import content from an existing vault — detects version, classifies notes, transforms frontmatter, rebuilds indexes
/om-project-archive Move completed project from work/active/ to work/archive/YYYY/, update all indexes — moves whole active/<Topic>/ clusters intact
/om-tidy Self-maintenance pass — acts on every hygiene flag: archives completed work, groups clusters, splits oversized notes, reports open loops. Never deletes, never commits

Usage Notes

Daily:

  • /om-standup replaces the manual session start — reads North Star, active work, tasks, git log
  • /om-dump processes freeform text and routes each piece to the correct note type and folder
  • /om-wrap-up is auto-triggered when you say "wrap up" — runs full session review

Editing & Synthesis:

  • /om-humanize calibrates against your actual writing samples, not a word blacklist. Detects context from frontmatter (review → corporate-confident, incident → precise, 1:1 → conversational). Run after drafting any note to make it sound human.
  • /om-weekly bridges standup and review brief — run at end of week for cross-session patterns, North Star drift, and uncaptured wins. Output is transient by default; offer to promote findings to brag doc or North Star.

Capture:

  • /om-capture-1on1 handles transcripts, raw notes, or summaries
  • /om-incident-capture takes Slack URLs and produces structured incident documentation
  • /om-slack-scan should be run AFTER /om-peer-scan to add context beyond code (leadership, communication, collaboration evidence)

Performance:

  • /om-peer-scan works best when launched as parallel agents (one per person)
  • /om-review-brief needs the private brief to exist first — it generates filtered versions from it

Maintenance:

  • /om-vault-audit should be run at the end of substantial sessions — catches stale indexes and mixed context
  • /om-vault-upgrade with no source upgrades this vault through ShardMind 0.2.1+: upgrade-plan.ts sorts every differing file into behind (take), ahead (keep; upstream candidate), merged or conflict before anything is applied. With a source, it imports content from an existing vault (older obsidian-mind or any Obsidian vault). Detects version, classifies notes, transforms frontmatter, fixes wikilinks, rebuilds indexes. Use --dry-run to preview.
  • /om-project-archive handles the active/ → archive/ move with index updates

Subagents

Agent Purpose Invoked by
brag-spotter Proactively finds uncaptured wins and competency gaps /om-wrap-up, /om-weekly
context-loader Loads all vault context about a person, project, incident, or concept Direct — "load context on X"
cross-linker Finds missing wikilinks, orphans, broken backlinks across the vault /om-vault-audit
people-profiler Bulk create/update person notes from Slack profiles /om-incident-capture
review-prep Aggregates all performance evidence for a given review period /om-review-brief
slack-archaeologist Full Slack reconstruction — reads every message, thread, profile, produces unified timeline /om-incident-capture
vault-librarian Deep vault maintenance — orphan detection, broken links, frontmatter validation, stale notes /om-vault-audit
review-fact-checker Verify every claim in a review draft against vault sources /om-self-review, /om-review-peer
vault-migrator Classify, transform, and migrate content from a source vault /om-vault-upgrade

Subagents run in isolated context windows via .claude/agents/. They don't pollute the main conversation.

Hooks

Hook When What
SessionStart On startup/resume QMD re-index, inject North Star, active work, recent changes, tasks, file listing
UserPromptSubmit Every message Classify content (decision, incident, 1:1, win, architecture, person, project update) and inject routing hints
PostToolUse After writing .md Validates frontmatter, checks for wikilinks
PreCompact Before context compaction Back up session transcript to thinking/session-logs/
Stop After every response Checklist + hygiene findings, once per session and again when they change, to the user and to the agent (which decides whether to act); ask the agent to run om-tidy to act

Semantic Search (QMD)

If QMD is installed (npm install -g @tobilu/qmd), the vault has semantic search. Every command takes --index <name>, where <name> is vault-manifest.json's qmd_index field when set and otherwise the vault folder name slugified:

  • qmd --index <name> query "..." — hybrid BM25 + vector + LLM reranking (best quality)
  • qmd --index <name> search "..." — fast BM25 keyword search
  • qmd --index <name> vsearch "..." — semantic vector search (exploratory)
  • qmd --index <name> update && qmd --index <name> embed — refresh index after bulk changes

SessionStart hook runs qmd --index <name> update automatically, reading the index name from the manifest. First-time setup on a fresh clone: node --experimental-strip-types .scripts/qmd-bootstrap.ts. See .claude/skills/qmd/SKILL.md for full reference, and [[Memories]] for the topics QMD is most often asked to find across the vault.

The om MCP Server (reaching this vault from another repo)

.claude/scripts/om-mcp.mjs exposes this vault over MCP so a session in another repository can use it. Registered in the consuming project's .mcp.json, never in this vault's own — the server refuses a memory write from a session running inside the vault, because such a memory would be scoped to the vault-as-a-project and reach only sessions that already read every note directly.

tool purpose
search semantic + keyword over the notes this vault serves
expand a note's links out and backlinks
recall durable lessons scoped to the calling repo, most specific first; explain: true reports what was withheld and why. A capture whose promoted: marker carries an anchor serves the promoted brain/ block instead of the capture body, so you read the corrected version
remember record a lesson that generalises beyond this repo
record_work record what happened here, filed where it belongs
reason judgement across several notes, by reading the vault with a second Claude session
health where memories live, which roots are exposed, is the index reachable

Two prompts you invoke yourself from the / menu: recall_topic, prior_art. They must be selected from the menu — typing the displayed label errors.

reason is the slow one. The rest answer without inference; this one spawns a session, so it takes seconds to minutes. Reach for it when search or recall returned the notes but not the answer — it seeds itself from search, so there is no need to search first. It runs on your own CLI default model unless reason.model pins one, nothing about it is capped, and every call is logged with its cost, turns, model and wall time. Answers land in .claude/om-reasoning/ marked confidence: inferred and are never auto-recorded as memories.

Debugging it: call health first. Every failure in this layer presents identically as "no results" — a renamed memory folder, a moved launcher, an index registered to a different vault, a session with no identity — and health is what tells them apart. A caller reported as ANONYMOUS means the client never completed the roots handshake, so only general-scope memories are visible.

Two guards worth knowing: memories are never served as ordinary notes (they carry their own declared scope), and a memory that would reach nobody is refused rather than silently widened to general.

Full setup, which notes the server serves, and the required repo-side snippet: .claude/rules/vault-from-another-repo.md. How the whole layer works — the four MCP surfaces, the identity handshake, a search call end to end, memory-visibility evaluation, the write path: ARCHITECTURE.md § Reaching the Vault From Another Repo.

Workflow: Weekly Review

  1. /om-weekly — synthesize the week's activity, check alignment, find patterns
  2. Promote any uncaptured wins to brag doc
  3. Update North Star if focus shifted
  4. /om-wrap-up — close the session cleanly

Workflow: Full Review Cycle Prep

  1. /om-review-brief manager — generate the manager context transfer doc
  2. /om-review-brief peers — generate the peer context transfer doc
  3. /om-peer-scan (parallel, one per peer) — deep scan each peer's PRs
  4. /om-slack-scan — scan relevant channels for your own evidence + peer context
  5. /om-capture-1on1 — capture the review 1:1 with your manager
  6. /om-vault-audit — tidy up after all the new data

Workflow: Project Ramp-Up

  1. /om-slack-scan — scan project channels for history and decisions
  2. /om-peer-scan (if needed) — understand what teammates have already built
  3. Create work note from gathered context
  4. /om-vault-audit — ensure everything links properly