### SessionStart Hook Example

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Demonstrates the structured context injected into an agent at the start of a session. This hook is automatically wired and fires on session events.

```bash
# Wired automatically via .claude/settings.json — fires on SessionStart event
# Agent receives structured context block:

## Session Context

### Date
Wednesday, 2026-04-30 (Week 18, Q2-2026)

### North Star (current goals)
[first 30 lines of brain/North Star.md]

### Brain Topics (read on demand)
- Gotchas — Known pitfalls and things that have gone wrong (empty)
- Key Decisions — Major technical and product decisions
- Memories — Index of memory topics
- Patterns — Recurring patterns observed across work
- Skills — Custom workflows and slash commands

### Recent Changes (last 48h)
abc1234 Add error monitoring to auth service
def5678 Update API contract draft

### Open Tasks
- [ ] Review auth PR with Sarah
- [ ] Draft incident RCA

### Active Work
- Auth Refactor.md
- API Contract.md

### Vault File Listing
./brain/Gotchas.md
./brain/Key Decisions.md
[... all .md files except .git, .obsidian, thinking, .claude ...]
```

--------------------------------

### Install Obsidian Mind via ShardMind

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Installs Obsidian Mind using ShardMind, a recommended method that includes a personalization wizard. Follow prompts for setup.

```bash
npm install -g shardmind
mkdir my-vault && cd my-vault
shardmind install github:breferrari/obsidian-mind
# Wizard prompts: your name, org, vault purpose, agents to include, QMD enabled
# Post-install hook personalizes brain/North Star.md and optionally bootstraps QMD
```

--------------------------------

### Install Obsidian Mind via Git Clone

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Installs Obsidian Mind by cloning the repository directly. Requires manual setup of Obsidian CLI and agent execution.

```bash
git clone https://github.com/breferrari/obsidian-mind.git my-vault
cd my-vault
# Open folder as an Obsidian vault
# Enable Obsidian CLI in Settings → General (requires Obsidian 1.12+)
# Run your agent: claude  OR  codex  OR  gemini
```

--------------------------------

### QMD Bootstrap Script for Index Setup

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

This script performs a one-time, idempotent setup for the QMD index. It registers the vault, attaches context, syncs ignore patterns, and builds the BM25 index and vector embeddings.

```bash
# Run from vault root:
node --experimental-strip-types scripts/qmd-bootstrap.ts

# Output:
# → Bootstrapping QMD index 'obsidian-mind'
# → Registering collection 'obsidian-mind' (pattern **/*.md)
# → Clearing previous context (if any)
# → Attaching vault context from manifest
# → Syncing ignore patterns from .obsidian/app.json
#   3 ignore pattern(s) synced from .obsidian/app.json
# → Indexing vault files
# → Generating embeddings
#
# ✓ QMD index 'obsidian-mind' ready. Test with:

```

--------------------------------

### Install and Bootstrap QMD

Source: https://github.com/breferrari/obsidian-mind/blob/main/README.md

Install the QMD package and run the bootstrap script to set up semantic search for your vault. This process is idempotent and configures the vault's index and context.

```bash
npm install -g @tobilu/qmd
node --experimental-strip-types scripts/qmd-bootstrap.ts
```

--------------------------------

### Install Obsidian Mind

Source: https://github.com/breferrari/obsidian-mind/blob/main/README.md

Install Obsidian Mind using either the `shardmind` command or by cloning the repository. Both methods result in the same vault.

```bash
shardmind install
```

```bash
git clone
```

--------------------------------

### Slash Command: /om-standup

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Initiates a morning kickoff by loading vault context and presenting a structured summary with suggested priorities. This command should be invoked at the start of any substantial work session.

```shell
/om-standup

# Agent reads:
# → Home.md (dashboard state)
# → brain/North Star.md (current goals)
# → work/Index.md (active projects)
# → obsidian daily:read (today's daily note)
# → obsidian tasks daily todo (open tasks)
# → git log --oneline --since="24 hours ago" --no-merges

# Output:
# **Yesterday**: Merged auth coordinator split (abc1234), updated API contract draft
# **Active Work**: Auth Refactor (blocked on error monitoring), API Contract (in review)
# **Open Tasks**: Review auth PR with Sarah, Draft incident RCA
# **North Star Alignment**: Auth work directly advances Q2 reliability goal
# **Suggested Focus**: Unblock error monitoring — Sarah flagged this at last 1:1
```

--------------------------------

### Session Start Hook Scripts

Source: https://github.com/breferrari/obsidian-mind/blob/main/CLAUDE.md

Example hook scripts for Obsidian. These scripts can automate tasks like session startup, message classification, and validation.

```typescript
session-start.ts
```

```typescript
classify-message.ts
```

```typescript
validate-write.ts
```

```typescript
pre-compact.ts
```

```typescript
stop-checklist.ts
```

```typescript
charcount.ts
```

--------------------------------

### QMD CLI Commands

Source: https://github.com/breferrari/obsidian-mind/blob/main/README.md

Example commands for interacting with QMD via the command-line interface. Use the `--index` flag to specify your vault's index name.

```bash
qmd --index obsidian-mind query "what did we decide about caching"
qmd --index obsidian-mind update
qmd --index obsidian-mind embed
```

--------------------------------

### Start Claude Code CLI

Source: https://github.com/breferrari/obsidian-mind/wiki/Obsidian-+-Claude-Code-Setup-Guide

Navigate to the cloned vault directory and start the Claude Code CLI. Claude will automatically detect the vault structure and load context.

```bash
cd obsidian-mind
claude
```

--------------------------------

### Note Frontmatter Schema Examples

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

These examples illustrate the required YAML frontmatter for different types of notes. They specify essential fields like date, description, and tags, along with type-specific fields enforced by validation hooks.

```yaml
# Work note (work/active/ or work/archive/YYYY/)
---
date: "2026-04-30"
description: "Refactoring the auth service to split coordinator role and add error monitoring before Q2 release."
project: auth-refactor
status: active          # active | completed | archived
quarter: Q2-2026
tags:
  - work-note
---

# Decision Record (work/active/ or work/)
---
date: "2026-04-30"
description: "Decision to defer Redis migration to Q2 and focus on API contract first."
status: proposed        # proposed | accepted | deprecated
tags:
  - decision
---
```

--------------------------------

### Bootstrap QMD Semantic Search

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Installs and bootstraps the QMD semantic search layer. This process is idempotent and may download large models on the first run.

```bash
npm install -g @tobilu/qmd
node --experimental-strip-types scripts/qmd-bootstrap.ts
# → Registers collection, attaches context, builds BM25 index + vector embeddings
# → Safe to re-run (idempotent)
# → First embed run downloads ~328MB embedding model + ~1.28GB reranker (for qmd query)
```

```bash
# Test the index:
qmd --index obsidian-mind query "what did we decide about caching"
qmd --index obsidian-mind search "auth refactor"   # BM25 only, no model download
```

--------------------------------

### Daily Standup Command

Source: https://github.com/breferrari/obsidian-mind/blob/main/README.md

Use the `/om-standup` command to get a summary of your current projects, tasks, and recent activities. It loads your North Star, active projects, open tasks, and recent git changes.

```bash
/om-standup
# → loads North Star, active projects, open tasks, recent git changes
# → "You have 2 active projects. The auth refactor is blocked on API contract.
#    Your 1:1 with Sarah is at 2pm — last time she flagged observability."
```

--------------------------------

### Install Obsidian Mind with ShardMind

Source: https://github.com/breferrari/obsidian-mind/blob/main/README.md

Use ShardMind to install the Obsidian Mind vault. This method includes a wizard for configuration and personalization. Ensure you are in a new directory before running the install command.

```bash
npm install -g shardmind
mkdir my-vault && cd my-vault
shardmind install github:breferrari/obsidian-mind
```

--------------------------------

### Gemini CLI Settings for Hooks

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Configure background scripts to run at specific points in the AI agent session lifecycle. Ensure Node.js is installed and the script paths are correct.

```json
{
  "hooks": {
    "SessionStart": "node --experimental-strip-types .claude/scripts/session-start.ts",
    "BeforeAgent": "node --experimental-strip-types .claude/scripts/classify-message.ts",
    "AfterTool": "node --experimental-strip-types .claude/scripts/validate-write.ts",
    "PreCompress": "node --experimental-strip-types .claude/scripts/pre-compact.ts"
  }
}
```

--------------------------------

### PR Title Format Example

Source: https://github.com/breferrari/obsidian-mind/blob/main/CONTRIBUTING.md

Use this format for PR titles to ensure consistency and proper changelog generation. Prefixes indicate the type of change.

```markdown
feat: add /om-review command
```

```markdown
fix: classify-message crash on empty input
```

```markdown
docs: update Japanese README with new commands
```

--------------------------------

### Classify Message Hook Example

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Demonstrates the classify-message hook, which processes user prompts from stdin to provide routing hints. It handles multiple languages and exits silently if no matches are found.

```bash
# Input (from agent harness via stdin):
echo '{"prompt": "Just had a 1:1 with Sarah. She praised the auth work and wants error monitoring before release.", "hook_event_name": "UserPromptSubmit"}' \
  | node --experimental-strip-types .claude/scripts/classify-message.ts

# Output (hookSpecificOutput envelope):
# Content classification hints (act on these if the user's message contains relevant info):
# - WIN detected — consider adding to perf/Brag Doc.md with a link to the evidence note
```

--------------------------------

### Adopt Existing Clone into ShardMind

Source: https://github.com/breferrari/obsidian-mind/blob/main/README.md

Use the `shardmind adopt` command to integrate an existing obsidian-mind clone into a managed v6 install. This process keeps your customizations and adds the necessary sidecar files.

```bash
npm install -g shardmind
shardmind adopt github:breferrari/obsidian-mind
```

--------------------------------

### QMD Bootstrap Process

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

This script outlines the steps involved in bootstrapping the QMD index. It covers reading manifest files, validating index names, adding collections and contexts, filtering ignored files, and performing initial updates for both BM25/FTS and vector embeddings.

```typescript
// What the bootstrap does (scripts/qmd-bootstrap.ts):
// 1. Reads vault-manifest.json for qmd_index, qmd_context, template fields
// 2. Validates qmd_index (alphanumeric + dot/dash/underscore, starts with alphanumeric)
// 3. qmd --index <name> collection add . <name> --pattern **/*.md
// 4. qmd --index <name> context rm qmd://<name>/   (idempotent — ignores "not found")
// 5. qmd --index <name> context add qmd://<name>/ "<qmd_context>"
// 6. Reads .obsidian/app.json userIgnoreFilters → writes to ~/.config/qmd/<name>.yml
// 7. qmd --index <name> update    (BM25/FTS index)
// 8. qmd --index <name> embed     (vector embeddings — downloads models on first run)
```

--------------------------------

### QMD CLI Equivalents for MCP Tools

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

These are the command-line interface equivalents for the MCP tools. They serve as a fallback when MCP is not registered, allowing direct interaction with the QMD index for querying, searching, and managing data.

```bash
# CLI equivalents (fallback when MCP not registered):
qmd --index obsidian-mind query "what did we decide about caching"
qmd --index obsidian-mind search "auth refactor"        # BM25 only, faster
qmd --index obsidian-mind vsearch "authentication"      # vector-only, no reranker
qmd --index obsidian-mind get "brain/Key Decisions.md"
qmd --index obsidian-mind update    # re-index after bulk edits
qmd --index obsidian-mind embed     # regenerate embeddings
```

--------------------------------

### Brag Spotter Subagent Output

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Example output format for the brag-spotter subagent, detailing uncaptured wins, competency gaps, and suggested brag entries.

```markdown
<!-- Invoked by /om-wrap-up and /om-weekly. Definition: .claude/agents/brag-spotter.md -->
<!-- tools: Read, Grep, Glob, Bash | model: sonnet | maxTurns: 20 -->

Output format:
**Uncaptured wins:**
- INC-2847 incident lead (high impact) — demonstrates Reliability competency
  Evidence: work/incidents/INC-2847 Auth Token Failure.md

**Competency gaps (thin evidence this quarter):**
- Mentorship — 0 evidence links from current quarter notes

**Suggested brag entries:**
> **Q2 2026 — Led P1 incident response for INC-2847**
> Root-caused auth connection pool exhaustion under load, coordinated fix,
> reduced MTTR to 39min. See [[INC-2847 Auth Token Failure]].
```

--------------------------------

### MCP Tools - In-Session Semantic Search

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

These are the in-session agent commands available when QMD is installed and the MCP server is registered. They provide typed access to QMD functionalities.

```APIDOC
## MCP Tools - In-Session Semantic Search

### Description
When QMD is installed and the MCP server is registered, the agent gains four typed tools that appear alongside Read and Edit in the tool menu. Preference order: MCP tools first (no `--index` argument needed — wrapper reads `qmd_index` from manifest), then `qmd` CLI fallback, then Grep/Glob/Read as last resort.

### Tools

#### `mcp__qmd__query`

##### Description
Semantic search - finds relevant notes by meaning, not keyword.

##### Usage
```typescript
mcp__qmd__query({ query: "what did we decide about caching" })
```

##### Returns
Ranked results with file paths, snippets, relevance scores.

#### `mcp__qmd__get`

##### Description
Retrieve a note by path.

##### Usage
```typescript
mcp__qmd__get({ path: "brain/Key Decisions.md" })
```

##### Returns
Full note content.

#### `mcp__qmd__multi_get`

##### Description
Retrieve multiple notes at once.

##### Usage
```typescript
mcp__qmd__multi_get({ paths: ["brain/Patterns.md", "brain/Gotchas.md"] })
```

##### Returns
Array of note contents.

#### `mcp__qmd__status`

##### Description
Check index status.

##### Usage
```typescript
mcp__qmd__status({})
```

##### Returns
Index name, collection count, last-updated timestamp.
```

--------------------------------

### Direct Invocation of SessionStart Hook

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Shows how to directly invoke the session-start script for testing or custom agent integration. It reads project directory from environment variables.

```typescript
// Direct invocation (for testing or custom agents):
// node --experimental-strip-types .claude/scripts/session-start.ts
// Reads CLAUDE_PROJECT_DIR / CODEX_PROJECT_DIR / GEMINI_PROJECT_DIR env vars
// Falls back to process.cwd()
```

--------------------------------

### Running Tests

Source: https://github.com/breferrari/obsidian-mind/blob/main/CONTRIBUTING.md

Execute tests by navigating to the scripts directory and running the npm test command. Tests are automatically triggered for PRs affecting the scripts.

```bash
cd .claude/scripts && npm test
```

--------------------------------

### QMD CLI Equivalents

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

These are the command-line interface equivalents for the QMD operations, used as a fallback when the MCP is not registered.

```APIDOC
## QMD CLI Equivalents

### Description
These are the command-line interface equivalents for the QMD operations, used as a fallback when the MCP is not registered.

### Commands

#### `qmd --index <name> query "<topic>"`

##### Description
Performs a semantic search for notes related to the specified topic.

#### `qmd --index <name> search "<keyword>"`

##### Description
Performs a BM25 keyword search, which is faster than semantic search.

#### `qmd --index <name> vsearch "<keyword>"`

##### Description
Performs a vector-only search without a reranker.

#### `qmd --index <name> get "<path>"`

##### Description
Retrieves the full content of a note by its path.

#### `qmd --index <name> update`

##### Description
Re-indexes notes after bulk edits.

#### `qmd --index <name> embed`

##### Description
Regenerates vector embeddings for notes.
```

--------------------------------

### Gemini CLI Configuration for Context Files

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Configure the Gemini CLI to recognize specific markdown files for context. This allows the CLI to read GEMINI.md and CLAUDE.md natively.

```json
// Gemini CLI — ~/.gemini/settings.json
// Read CLAUDE.md natively:
{ "context": { "fileName": ["GEMINI.md", "CLAUDE.md"] } }
```

--------------------------------

### Slash Command: /om-dump

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Captures freeform natural language input, classifies it, and routes it to the correct vault notes with proper templates and wikilinks. It searches for existing notes before creating new ones.

```shell
/om-dump Just had a 1:1 with Sarah. She's happy with the auth work but wants
error monitoring before release. Tom mentioned the cache migration is deferred
to Q2 — we decided to focus on the API contract first. Win: Sarah praised
the auth architecture in front of the whole team.

# Agent:
# 1. Classifies: 1:1 content, decision, win, person context
# 2. Searches QMD for existing Sarah, Tom, auth refactor notes
# 3. Creates work/1-1/Sarah Chen 2026-04-30.md with full frontmatter + wikilinks
# 4. Creates Decision Record: "Defer Redis migration to Q2"
# 5. Appends to perf/Brag Doc.md: "Auth architecture praised by manager"
# 6. Updates org/people/Sarah Chen.md with meeting context
# 7. Updates work/active/Auth Refactor.md with error monitoring task

# Summary:
# → Created: work/1-1/Sarah Chen 2026-04-30.md
# → Created: work/active/Defer Redis Migration Decision.md
# → Updated: perf/Brag Doc.md (Q2 2026 section)
# → Updated: org/people/Sarah Chen.md
```

--------------------------------

### Dry Run Vault Migration

Source: https://github.com/breferrari/obsidian-mind/blob/main/README.md

Use the `--dry-run` flag with the `/om-vault-upgrade` command to preview the migration plan without executing any changes. This is useful for understanding what will be copied, transformed, and skipped.

```bash
# 1. Clone the latest obsidian-mind
git clone https://github.com/breferrari/obsidian-mind.git ~/new-vault

# 2. Open it in your agent
cd ~/new-vault && claude   # or codex, or gemini

# 3. Run the upgrade pointing to your old vault
/om-vault-upgrade --dry-run ~/my-old-vault
```

--------------------------------

### QMD Integration Flowchart

Source: https://github.com/breferrari/obsidian-mind/blob/main/ARCHITECTURE.md

Visual representation of how agent tools and index maintenance processes interact with the QMD system and its storage.

```mermaid
flowchart TB
    subgraph Reader["Agent tools"]
        MCP_get["mcp__qmd__get"]
        MCP_query["mcp__qmd__query"]
        MCP_multi["mcp__qmd__multi_get"]
        MCP_status["mcp__qmd__status"]
    end
    subgraph Writer["Index maintenance"]
        Boot["qmd-bootstrap.ts<br/>(one-time)"]
        Session["session-start.ts<br/>(per session)"]
        Refresh["qmd-refresh-run.ts<br/>(per write/stop/pre-compact,<br/>debounced, detached)"]
    end
    Wrapper["qmd-mcp.mjs<br/>(reads qmd_index from manifest)"]
    CLI["qmd CLI<br/>(--index &lt;name&gt; from manifest)"]
    Store[("Named SQLite store<br/>+ embeddings")]

    MCP_get --> Wrapper
    MCP_query --> Wrapper
    MCP_multi --> Wrapper
    MCP_status --> Wrapper
    Wrapper --> Store
    Boot --> CLI
    Session --> CLI
    Refresh --> CLI
    CLI --> Store
```

--------------------------------

### Pre-Compact Hook Script

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

This script acts as a PreCompact hook, backing up the session transcript before context compaction. It also triggers a debounced QMD refresh and manages backup retention.

```bash
# Input (from agent harness via stdin before compaction):
echo '{"transcript_path": "/tmp/session-abc123.jsonl", "trigger": "manual"}' \
  | node --experimental-strip-types .claude/scripts/pre-compact.ts

# Effect:
# → Copies transcript to thinking/session-logs/session_manual_20260430_143022.jsonl
# → Prunes backups older than the 30 most recent
# → Triggers debounced QMD refresh (30s debounce window, detached worker)
# → Exit 0 always

# Backup naming: session_<trigger>_<YYYYMMDD_HHMMSS>.jsonl

```

--------------------------------

### Ignore List Synchronization Flowchart

Source: https://github.com/breferrari/obsidian-mind/blob/main/ARCHITECTURE.md

Diagram illustrating how the ignore list from Obsidian's configuration is propagated to QMD's index.

```mermaid
flowchart LR
    App[".obsidian/app.json<br/>userIgnoreFilters"]
    Boot["qmd-bootstrap.ts"]
    YAML["~/.config/qmd/&lt;index&gt;.yml<br/>collections.&lt;name&gt;.ignore"]
    Obs["Obsidian<br/>(search, graph, switcher)"]
    QmdUpdate["qmd update<br/>(every invocation)"]

    App --> Obs
    App --> Boot
    Boot --> YAML
    YAML --> QmdUpdate
```

--------------------------------

### Lifecycle Hooks Sequence Diagram

Source: https://github.com/breferrari/obsidian-mind/blob/main/ARCHITECTURE.md

This diagram illustrates the sequence of events and interactions between the user, agent, hooks, vault, and QMD during a session, highlighting data flow and hook responsibilities.

```mermaid
sequenceDiagram
    participant User
    participant Agent
    participant Hooks
    participant Vault
    participant QMD

    User->>Agent: start session
    Agent->>Hooks: SessionStart
    Hooks->>Vault: read North Star, git log, tasks, file listing
    Hooks->>QMD: re-index (async)
    Hooks-->>Agent: inject ~2K tokens of context

    loop each user message
        User->>Agent: prompt
        Agent->>Hooks: UserPromptSubmit
        Hooks->>Hooks: classify (decision, incident, win, 1:1, ...)
        Hooks-->>Agent: routing hints
    end

    loop each Write/Edit to .md
        Agent->>Hooks: PostToolUse
        Hooks->>Hooks: validate frontmatter + wikilinks
        Hooks->>QMD: debounced refresh (detached)
        Hooks-->>Agent: warnings if invalid
    end

    Note over Agent,Hooks: if context fills
    Agent->>Hooks: PreCompact
    Hooks->>Vault: back up transcript to thinking/session-logs/

    User->>Agent: end session
    Agent->>Hooks: Stop
    Hooks-->>Agent: wrap-up checklist reminder
```

--------------------------------

### Vault-First Memory Diagram

Source: https://github.com/breferrari/obsidian-mind/blob/main/ARCHITECTURE.md

Visualizes the two memory systems: ephemeral (auto-loaded index in ~/.claude/) and durable (git-tracked vault in brain/). It shows how the SessionStart hook and the auto-loaded index interact with the vault's topic notes.

```mermaid
flowchart LR
    SessionStart["SessionStart hook<br/>(.claude/scripts/session-start.ts)"]
    subgraph Ephemeral["~/.claude/ (not git-tracked)"]
        MemIndex["MEMORY.md<br/>(auto-loaded index)"]
    end
    subgraph Durable["Vault (git-tracked)"]
        NorthStar["brain/North Star.md"]
        BrainIdx["brain/Memories.md<br/>(topic index)"]
        Gotchas["brain/Gotchas.md"]
        Patterns["brain/Patterns.md"]
        Decisions["brain/Key Decisions.md"]
    end

    SessionStart ==>|reads every session| NorthStar
    MemIndex -->|points at| BrainIdx
    MemIndex -->|points at| Gotchas
    MemIndex -->|points at| Patterns
    MemIndex -->|points at| Decisions
    BrainIdx --> Gotchas
    BrainIdx --> Patterns
    BrainIdx --> Decisions
```

--------------------------------

### Claude Agent Configuration - settings.json

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Configuration for Claude agent hooks, defining commands to run at different stages of a session like startup, prompt submission, and tool use.

```json
{
  "hooks": {
    "SessionStart": [{
      "matcher": "startup|resume|clear|compact",
      "hooks": [{
        "type": "command",
        "command": "node --experimental-strip-types \"${CLAUDE_PROJECT_DIR:-.}/.claude/scripts/session-start.ts\"",
        "timeout": 30
      }]
    }],
    "UserPromptSubmit": [{
      "hooks": [{
        "type": "command",
        "command": "node --experimental-strip-types \"${CLAUDE_PROJECT_DIR:-.}/.claude/scripts/classify-message.ts\"",
        "timeout": 15
      }]
    }],
    "PostToolUse": [{
      "matcher": "Write|Edit",
      "hooks": [
        {
          "type": "command",
          "command": "node --experimental-strip-types \"${CLAUDE_PROJECT_DIR:-.}/.claude/scripts/validate-write.ts\"",
          "timeout": 15
        },
        {
          "type": "command",
          "command": "node --experimental-strip-types \"${CLAUDE_PROJECT_DIR:-.}/.claude/scripts/qmd-refresh.ts\"",
          "timeout": 15
        }
      ]
    }],
    "PreCompact": [{
      "hooks": [{
        "type": "command",
        "command": "node --experimental-strip-types \"${CLAUDE_PROJECT_DIR:-.}/.claude/scripts/pre-compact.ts\"",
        "timeout": 30
      }]
    }],
    "Stop": [{
      "hooks": [{
        "type": "command",
        "command": "node --experimental-strip-types \"${CLAUDE_PROJECT_DIR:-.}/.claude/scripts/stop-checklist.ts\"",
        "timeout": 5
      }]
    }]
  }
}
```

--------------------------------

### Multi-Agent Portability Diagram

Source: https://github.com/breferrari/obsidian-mind/blob/main/ARCHITECTURE.md

Illustrates how different agents (Claude Code, Codex, Gemini) map their unique event names to shared scripts via individual configuration files. Each agent also has its own operating manual.

```mermaid
flowchart TB
    subgraph Configs["Per-agent config (event name mapping)"]
        Claude[".claude/settings.json<br/>Claude Code"]
        Codex[".codex/hooks.json<br/>Codex CLI"]
        Gemini[".gemini/settings.json<br/>Gemini CLI"]
    end
    subgraph Shared[".claude/scripts/ (shared core)"]
        S1["session-start.ts"]
        S2["classify-message.ts"]
        S3["validate-write.ts"]
        S4["qmd-refresh.ts"]
        S5["pre-compact.ts"]
        S6["stop-checklist.ts"]
    end
    subgraph Manuals["Operating manuals"]
        CM["CLAUDE.md"]
        AM["AGENTS.md"]
        GM["GEMINI.md"]
    end

    Claude --> Shared
    Codex --> Shared
    Gemini --> Shared
    Claude -.reads.-> CM
    Codex -.reads.-> AM
    Gemini -.reads.-> GM
```

--------------------------------

### Perform Session Wrap-up and Review

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Conducts a full session review, verifying note quality, checking index consistency, finding orphans, identifying archivable projects, surfacing patterns, and running the brag-spotter subagent. This command can be triggered automatically by saying 'wrap up'.

```shell
/om-wrap-up
```

--------------------------------

### Obsidian CLI Commands

Source: https://github.com/breferrari/obsidian-mind/blob/main/CLAUDE.md

Use these commands when Obsidian is running for vault-aware operations. 'file=' resolves like a wikilink, 'path=' uses the exact path. Use 'silent' to prevent files from opening. Run 'obsidian help' for a full reference.

```bash
obsidian read file="Note Name"
```

```bash
obsidian create name="Name" content="..." silent
```

```bash
obsidian append file="Name" content="..."
```

```bash
obsidian search query="text" limit=10
```

```bash
obsidian backlinks file="Name"
```

```bash
obsidian tags sort=count counts
```

```bash
obsidian tasks daily todo
```

```bash
obsidian daily:read
```

```bash
obsidian property:set name="status" value="done" file="Name"
```

```bash
obsidian orphans
```

--------------------------------

### Codex CLI Configuration for Project Docs

Source: https://context7.com/breferrari/obsidian-mind/llms.txt

Specify fallback filenames for project documentation when using the Codex CLI. This ensures CLAUDE.md is read natively.

```toml
# Codex CLI — ~/.codex/config.toml
# Read CLAUDE.md natively:
project_doc_fallback_filenames = ["CLAUDE.md"]
```

--------------------------------

### QMD Index Name Coexistence

Source: https://github.com/breferrari/obsidian-mind/blob/main/ARCHITECTURE.md

The `qmd_index` field in the manifest ensures that multiple QMD-indexed projects can coexist on the same machine without data collision. Changing this field on the next bootstrap creates an isolated store.

```mermaid
flowchart LR
    Manifest["vault-manifest.json<br/>qmd_index: obsidian-mind"]
    SessionStart["SessionStart hook<br/>(session-start.ts)"]
    MCP[".mcp.json wrapper<br/>(qmd-mcp.mjs)"]
    Refresh["Mid-session refresh<br/>(PostToolUse / Stop / PreCompact<br/>→ qmd-refresh-run.ts worker)"]
    Store[("QMD SQLite store<br/>(named: obsidian-mind)")]

    Manifest --> SessionStart
    Manifest --> MCP
    Manifest --> Refresh
    SessionStart -->|re-index at startup| Store
    Refresh -->|debounced refresh| Store
    MCP -->|search tools| Store
```