total-recall-brain 3.13.0 → 3.13.2

package/README.md

@@ -1,173 +1,156 @@
-# Total Recall — SSSS Sovereign AI OS
+# Total Recall
 
-[![Version](https://img.shields.io/badge/version-3.3.0-indigo.svg)](package.json)
+[![Version](https://img.shields.io/badge/version-3.13.0-indigo.svg)](package.json)
 [![License](https://img.shields.io/badge/license-MIT-emerald.svg)](LICENSE)
-[![Continuous Integration](https://img.shields.io/badge/CI-passing-teal.svg)](.github/workflows/test.yml)
 
-> **A database-free, filesystem-native AI memory layer.** Works in any IDE. Self-hosts on any VM. Costs nothing to store.
+**Portable personal memory for any IDE** — filesystem-native, database-free, open source.
 
-Total Recall is the open-source reference implementation of the **SSSS (Structured Semantic Syntax System)** — a specification for storing AI memory, skills, and workflows as plain Markdown files instead of opaque, hosted databases.
+Total Recall stores your rules, preferences, facts, and project knowledge as plain Markdown (SSSS). It compiles them into IDE instruction surfaces, runs a dream consolidation cycle, and lets agents enqueue background tasks. Host apps and product repos are equal implementations: **nothing is hard-coded to a specific codebase.**
 
 ```text
-your AI brain = ~/.agent/skills/total-recall/memory-vault/*.md
+write   →  remember / session ingest
+sleep   →  dream  (consolidate, conflict, compile, prune)
+read    →  recall + compiled surfaces
+async   →  daemon task queue (agents may enqueue anything under policy)
 ```
 
-No database overhead. No vendor lock-in. Your thoughts, on your machine.
+```text
+your brain  =  ~/.agent/skills/total-recall/
+your project  =  <repo>/.agent/skills/total-recall/
+```
 
 ---
 
-## 🚀 Choose Your Setup
-
-### 🖥️ 1. IDE Memory Only (2 minutes)
+## Quick start (IDE memory, ~2 minutes)
 
-Add persistent, highly contextual memory to **Claude Code**, **Cursor**, **Codex CLI**, or any IDE agent in your active workspace:
+No always-on LLM and no research daemon required.
 
 ```bash
-cd ~/my-project
-npx total-recall init                   # provision global brain and local project VFS shims
-npx total-recall connect claude-code     # symlink CLAUDE.md → INSTRUCTIONS.md
+cd /path/to/any-project
+npx total-recall init --project    # project brain + openwiki scaffold
+npx total-recall connect claude-code   # or: cursor | codex | gemini | aider | obsidian | http-api
+npx total-recall remember preference "Prefer clear, short answers."
+npx total-recall recall "short answers"
+npx total-recall compile           # rebuild INSTRUCTIONS.md inject blocks
 ```
 
-Supported IDEs and surfaces:
+Injected rules sit only between:
 
-| Client Command | Connection Mode | Targeted Projection Target |
-| :--- | :--- | :--- |
-| `npx total-recall connect claude-code` | Symlink | Symlinks `CLAUDE.md` to the compiled instruction surface |
-| `npx total-recall connect codex` | Symlink | Symlinks `AGENTS.md` to the compiled instruction surface |
-| `npx total-recall connect antigravity` | Symlink | Symlinks `AGENTS.md` to the compiled instruction surface |
-| `npx total-recall connect gemini` | Symlink | Symlinks `GEMINI.md` to the compiled instruction surface |
-| `npx total-recall connect cursor` | File Projection | Writes/updates `.cursor/rules/total-recall.mdc` with frontmatter rules |
-| `npx total-recall connect aider` | File Projection | Writes/updates `.aider.rules.md` file rules |
+```text
+<!-- BEGIN INJECTED MEMORY: do not edit by hand; rebuilt by total-recall surface -->
+…
+<!-- END INJECTED MEMORY -->
+```
 
-> [!NOTE]
-> **Non-Destructive Integration:**
-> Your existing custom rules are **never overwritten**. Total Recall operates exclusively inside the clearly-marked comment bounds:  
-> `<!-- BEGIN INJECTED MEMORY: do not edit by hand; rebuilt by total-recall surface -->`  
-> `...`  
-> `<!-- END INJECTED MEMORY -->`
+Existing local rules outside those markers are never clobbered.
 
 ---
 
-### 🧠 2. Full Sovereign Brain VM (10 minutes)
+## Core workflows
 
-Deploy the full autonomous stack — a secure OpenAI-compatible REST server, dynamic React dashboard, continuous Dream Cycle daemon, and local session relays — on your own VM.
+| Workflow | Commands |
+|----------|----------|
+| **Remember / recall** | `remember`, `forget`, `recall` |
+| **Dream (sleep)** | `dream` — or daemon system dream on empty queue |
+| **Defer work** | `task add "<intent>"` then `daemon start` |
+| **Skills across repos** | `skill track <path>`, `skill deploy`, `skill sync` |
+| **Secrets (not vault)** | `secret set\|list\|rotate\|usage\|check-surfaces` |
+| **Connect an IDE** | `connect <client>` |
 
-**Prerequisites:**
-- A standard Linux or macOS server/workstation.
-- Node.js ≥ 20.
-- A primary API Key (e.g. `GOOGLE_API_KEY` for high-speed Google `gemini-embedding-2` vectors, featuring OpenAI fallback).
+### Multi-repo skills (any path you choose)
 
 ```bash
-# Start the interactive configuration wizard:
-npx total-recall init
+npx total-recall skill track /path/to/any-app
+npx total-recall skill sync --repo /path/a --repo /path/b
+# or: export TR_SYNC_REPOS="/path/a:/path/b"
 ```
 
-The interactive wizard provisions the configuration, registers authorized Personal Access Tokens (PATs), configures secure Caddy reverse-proxying, and prompts you to choose your **Selectable UI Deployment Location**:
+Roots come only from: project registry, install map, `TR_SYNC_REPOS`, CLI `--repo`, and cwd when it looks like a project.
 
-| Mode | Target URL | Primary Target |
-| :--- | :--- | :--- |
-| **Local Bind** | `http://127.0.0.1:3000` | Local operations. Binds strictly to your loopback interface. |
-| **Quick Tunnel** | `https://*.trycloudflare.com` | Spawns a background Cloudflare tunnel for zero-configuration public URL access. |
-| **Named Tunnel** | `https://your-tunnel.domain.com` | Hooks up securely to a pre-registered Cloudflare tunnel daemon. |
-| **Custom Domain** | `https://your-domain.com` | Hooks up to your public DNS, reverse-proxied and secured by Caddy TLS. |
+### Secrets
 
-To start the server daemon, run:
 ```bash
-npm start
+npx total-recall secret set some_api_key "$KEY" --provider example
+npx total-recall secret list                 # metadata only
+npx total-recall secret check-surfaces       # fail if values leaked into surfaces
 ```
-*Vite automatically installs and builds the beautiful Glassmorphic React SPA on first boot, serving the dashboard dynamically.*
-
----
 
-### 📔 3. Obsidian Integration
+Optional AES: `TR_SECRETS_PASSWORD`. Values never belong in vault markdown or openwiki.
 
-Visualize your entire memory vault, rules network, and daily logs directly inside Obsidian with native backlinks and graphs:
+### Optional background daemon
 
 ```bash
-npx total-recall connect obsidian
+npx total-recall daemon start    # idle invent OFF unless TR_IDLE_TASKS=1
+npx total-recall task add "Extract decisions from last session" --cap vault:write
 ```
-*Auto-detects your vault path on macOS (or pass `--vault ~/path/to/vault` on Linux/other), symlinking your memory vault and installing interactive Dataview dashboard trackers.*
 
 ---
 
-### 💬 4. UltraChat Integration
-
-UltraChat integrates securely with your brain via our OpenAI-compatible chat API, injecting active memory instructions into system prompts on the fly:
+## Dual-layer brain
 
-```bash
-npx total-recall connect ultrachat --brain https://your-server.com --token YOUR_PAT
+```text
+GLOBAL   ~/.agent/skills/total-recall/     identity, shared preferences
+PROJECT  <repo>/.agent/skills/total-recall/  repo facts, decisions, openwiki
 ```
-*Generates the OpenAI-compatible configuration parameters (baseURL, model, token auth header).*
+
+CLI flags: `--global` / `--project`. Project wins on slug conflicts when both apply.
 
 ---
 
-## 🧠 Dual-Layer Brain Architecture
+## CLI inventory
 
-Total Recall partitions your sovereign brain memory into two virtual layers:
+Commands are classified as **core** (default product path), **optional** (power features), or **legacy** (still shipped, not the focus).
 
-```
-                      ┌────────────────────────┐
-                      │      GLOBAL LAYER      │
-                      │  ~/.agent/skills/tr/   │
-                      └───────────┬────────────┘
-                                  │
-                                  ▼ Overridden by
-                      ┌────────────────────────┐
-                      │     PROJECT LAYER      │
-                      │  <repo>/.agent/skills/ │
-                      └────────────────────────┘
-```
+Full table: [`docs/reference/CLI_INVENTORY.md`](docs/reference/CLI_INVENTORY.md)
 
-1. **Global Layer** (`~/.agent/skills/total-recall/`): Contains global user preferences, master API keys, global logs, and the baseline identity context (`SOUL.md`, `USER.md`).
-2. **Project Layer** (`<repo>/.agent/skills/total-recall/`): Contains local repository facts, project-specific habits, and specialized research agendas.
+| Tier | Examples |
+|------|----------|
+| **Core** | `init`, `connect`, `remember`, `forget`, `recall`, `compile`, `dream`, `task`, `daemon`, `skill`, `secret`, `brain`, `status`, `doctor` |
+| **Optional** | `research`, `relay`, `deploy`, `setup`, `backup`/`restore`, `chat`, `map`, `export`/`import`, `ingest` |
+| **Legacy / niche** | `collab`, `friction`, `upgrade`, `migrate`, `snapshot`, `command` |
 
-*When compiling rules or searching, project-level nodes take precedence and override global nodes on slug collisions. CLI operations like `remember`, `recall`, `research`, `backup`, and `restore` accept `--global` and `--project` flags to steer memories to the right layer.*
+Default install story: **init → connect → remember/recall → dream**. No LLM required.
 
 ---
 
-## ⚡ CLI Command Reference
-
-| Command | Description |
-| :--- | :--- |
-| `npx total-recall init` | Initialize the global or local project VFS memory schemas. |
-| `npx total-recall connect <client>` | Wire a local IDE, Obsidian vault, or UltraChat client to your brain. |
-| `npx total-recall deploy` | Configure startup services, Caddy proxy servers, and platform auto-starts. |
-| `npx total-recall compile` | Rebuild `INSTRUCTIONS.md` from merged global & project memory nodes. |
-| `npx total-recall dream` | Trigger a manual Dream Cycle (GC, pattern extraction, decay consolidation). |
-| `npx total-recall lint` | Validate vault nodes against the SSSS v2 frontmatter Zod schema. |
-| `npx total-recall backup` | Create password-encrypted tarballs, push git backups, or sync Obsidian. |
-| `npx total-recall restore <file>` | Restore your sovereign VFS from an encrypted tarball backup. |
-| `npx total-recall status` | Show brain health summary, connected client registries, and daemon status. |
-| `npx total-recall generate-pat` | Provision granular, label-based Personal Access Tokens (PATs) for auth. |
-| `npx total-recall daemon` | Manage the background Dream Cycle daemon (start / stop / status). |
-| `npx total-recall relay` | Manage background local session-sync relays watching IDE log files. |
-| `npx total-recall config` | Read or update system settings, Allowed Origins, or USD budget caps dynamically. |
-| `npx total-recall skill` | Search, install, security-audit, and remove packages from the skills.sh registry. |
-| `npx total-recall uninstall` | Completely stop background services and wipe Total Recall from the system. |
-| `npx total-recall friction` | Inspect watchdog telemetry logs to identify workflow latency bottlenecks. |
+## Optional full server
 
----
+For a local REST API + dashboard (not required for IDE memory):
 
-## 🔒 Security, Sandboxing, & Safety
+```bash
+npx total-recall start
+# or: npm start
+```
 
-- **硬 Hardened VM Sandbox:** All subagent script executions run inside an isolated vm sandbox environment (`sandbox.mjs`) featuring POSIX namespace restrictions, strict memory/CPU bounds, and a command execution whitelist. The sandbox is **disabled by default** (`security.yml.sandbox.enabled: false`) for maximum security.
-- **Cost Watchdog & Limits:** Daily and weekly maximum cost thresholds (`config/budget.yml`) are enforced dynamically via a cost supervisor in `runtime.mjs` that instantly suspends subagent dispatches if budget caps are exceeded.
-- **OWASP-aligned scrypt Security:** Master credentials and API keys are AES-256-GCM encrypted under `secrets.enc` with OWASP-aligned scrypt key derivation.
-- **Privacy Redactions:** Memory nodes marked with `privacy: local_only` are automatically stripped before any external API dispatches or reasoning agent calls.
+Deploy/tunnels are optional (`deploy`, `setup`). Configure only what you need.
 
 ---
 
-## 🧹 Service Uninstallation
+## Architecture sketch
+
+- **SSSS vault** — Markdown + YAML frontmatter (filesystem SSOT)
+- **Openwiki** — human/agent long-form docs (ships with init)
+- **Surfaces** — compiled inject blocks for IDEs
+- **Dream** — consolidation cycle (deterministic first)
+- **Daemon tasks** — open envelope + capability policy
+- **Secrets** — separate store from vault
 
-To cleanly stop background daemons, unload platform agents (macOS launchd plists / Linux systemd services), remove symlink shims from all workspaces, and purge global configs:
+Host apps consume TR the same way: PAT + brain URL (`connect http-api`) or local files.
+
+---
+
+## Uninstall
 
 ```bash
 npx total-recall uninstall
 ```
-> [!IMPORTANT]
-> **Safety First:**
-> The uninstaller automatically detects version-controlled files, **preserving** local workspace `.agent/skills/` and `.agent/memory-vault/` directories inside active git project repositories to ensure you never lose custom rules and memories.
+
+Project brains inside git repos are preserved when possible so you do not lose custom rules and memories.
 
 ---
 
-*Sovereign. Database-free. Yours.*
+## License
+
+MIT — see [LICENSE](LICENSE).
+
+*Portable memory. Your files. Any IDE.*

package/bin/antigravity.mjs

@@ -1,20 +1,17 @@
 #!/usr/bin/env node
-// Antigravity CLI Agent Wrapper — routes through the running Antigravity IDE
-// via agentapi, using the user's logged-in Google account.
 
 import fs from 'node:fs';
 import path from 'node:path';
 import os from 'node:os';
-import { spawnSync } from 'node:child_process';
 
 const args = process.argv.slice(2);
 let prompt = '';
 let outputJson = false;
-let requestedModel = process.env.TR_CLI_MODEL || 'flash';
+let requestedModel = process.env.TR_CLI_MODEL || 'default';
 
 if (args.includes('-h') || args.includes('--help')) {
   console.log(`
-Antigravity CLI Agent Wrapper — Headless via agentapi (logged-in account)
+Antigravity CLI Agent Wrapper — Headless Zero-Dependency Cognitive Agent
 
 Usage:
   antigravity [options] [prompt]
@@ -22,15 +19,15 @@ Usage:
 
 Options:
   -p, --prompt    The text prompt to generate content for
-  -m, --model     Model: flash_lite, flash, pro (default: flash)
-  -o json         Output JSON response payload
-  --json          Output JSON response payload
+  -m, --model     Gemini API model identifier or alias (default: default / gemini-3.5-flash)
+  -o, --json      Output dynamic JSON response payload
+  --json          Output dynamic JSON response payload
   -h, --help      Show this help information
 `);
   process.exit(0);
 }
 
-// Ignored flags for compatibility with tr-cli-agents dispatch
+// Ignored flags for compat with tr-cli-agents dispatch
 const IGNORED = new Set(['--sandbox=false', '--yolo']);
 
 for (let i = 0; i < args.length; i++) {
@@ -42,7 +39,6 @@ for (let i = 0; i < args.length; i++) {
 }
 
 if (!prompt && args.length > 0) {
-  // Last non-flag arg is the prompt
   for (let i = args.length - 1; i >= 0; i--) {
     if (!args[i].startsWith('-') && !IGNORED.has(args[i])) { prompt = args[i]; break; }
   }
@@ -53,50 +49,83 @@ if (!prompt) {
   process.exit(1);
 }
 
-// Map model aliases
-const modelMap = {
-  default: 'flash', gemini: 'flash', 'gemini-flash': 'flash',
-  'gemini-3.5-flash': 'flash', '3.5-flash': 'flash',
-  pro: 'pro', 'gemini-pro': 'pro', '3.1-pro': 'pro',
-  lite: 'flash_lite', 'flash-lite': 'flash_lite',
-};
-const model = modelMap[requestedModel] || requestedModel;
-
-// Spawn agentapi new-conversation and capture the response
-const result = spawnSync('agentapi', [
-  'new-conversation',
-  `--model=${model}`,
-  '--title=antigravity-cli',
-  prompt,
-], {
-  encoding: 'utf8',
-  timeout: 300_000,
-  maxBuffer: 10 * 1024 * 1024,
-  env: process.env,
-});
-
-if (result.status !== 0) {
-  const err = result.stderr || result.stdout || 'agentapi failed';
-  console.error('Error:', err);
-  process.exit(1);
+// Load GOOGLE_API_KEY from env or .env files
+if (!process.env.GOOGLE_API_KEY) {
+  try {
+    let dir = process.cwd();
+    while (dir !== path.dirname(dir)) {
+      const envPath = path.join(dir, '.env');
+      if (fs.existsSync(envPath)) {
+        const content = fs.readFileSync(envPath, 'utf8');
+        for (const line of content.split('\n')) {
+          const match = line.match(/^\s*GOOGLE_API_KEY\s*=\s*(["']?)(.*?)\1\s*$/);
+          if (match) { process.env.GOOGLE_API_KEY = match[2]; break; }
+        }
+      }
+      if (process.env.GOOGLE_API_KEY) break;
+      dir = path.dirname(dir);
+    }
+  } catch {}
 }
 
-const raw = result.stdout?.trim() || '';
-
-// Parse agentapi response
-let text = raw;
-try {
-  const parsed = JSON.parse(raw);
-  // agentapi returns { response: { conversationId, content } }
-  if (parsed.response?.content) text = parsed.response.content;
-  else if (parsed.content) text = parsed.content;
-  else if (parsed.response) text = typeof parsed.response === 'string' ? parsed.response : JSON.stringify(parsed.response);
-} catch {
-  // Not JSON, use raw output
+const apiKey = process.env.GOOGLE_API_KEY || '';
+if (!apiKey) {
+  console.error('Error: GOOGLE_API_KEY not set.');
+  process.exit(1);
 }
 
-if (outputJson) {
-  console.log(JSON.stringify({ response: text }));
-} else {
-  console.log(text);
+// Model alias resolution
+const aliases = {
+  gemini: 'gemini-3.5-flash', default: 'gemini-3.5-flash',
+  flash: 'gemini-3.5-flash', 'gemini-flash': 'gemini-3.5-flash', '3.5-flash': 'gemini-3.5-flash',
+  pro: 'gemini-3.1-pro-preview', 'gemini-pro': 'gemini-3.1-pro-preview', '3.1-pro': 'gemini-3.1-pro-preview',
+};
+const resolvedModel = aliases[requestedModel] || requestedModel;
+
+async function main() {
+  try {
+    const url = `https://generativelanguage.googleapis.com/v1beta/models/${resolvedModel}:generateContent?key=${apiKey}`;
+
+    const res = await fetch(url, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify({
+        contents: [{ parts: [{ text: prompt }] }]
+      })
+    });
+
+    if (!res.ok) {
+      const errText = await res.text();
+      throw new Error(`Gemini API returned status ${res.status}: ${errText}`);
+    }
+
+    const data = await res.json();
+    const text = data.candidates?.[0]?.content?.parts?.[0]?.text || '';
+
+    // Save usage data for budget safety
+    try {
+      const promptTokens = data.usageMetadata?.promptTokenCount || 0;
+      const candidatesTokens = data.usageMetadata?.candidatesTokenCount || 0;
+      if (promptTokens > 0 || candidatesTokens > 0) {
+        const tmpDir = path.join(os.homedir(), '.gemini', 'tmp', 'antigravity', 'chats');
+        if (!fs.existsSync(tmpDir)) fs.mkdirSync(tmpDir, { recursive: true });
+        const usageLine = JSON.stringify({
+          timestamp: new Date().toISOString(),
+          tokens: { input: promptTokens, output: candidatesTokens }
+        });
+        fs.appendFileSync(path.join(tmpDir, 'usage.jsonl'), usageLine + '\n', 'utf8');
+      }
+    } catch {}
+
+    if (outputJson) {
+      console.log(JSON.stringify({ response: text }));
+    } else {
+      console.log(text);
+    }
+  } catch (err) {
+    console.error('Error calling Gemini API:', err.message);
+    process.exit(1);
+  }
 }
+
+main();

package/bin/total-recall.mjs

@@ -59,6 +59,7 @@ const COMMANDS = {
   recall:   'recall.mjs',
   search:   'recall.mjs',
   research: 'research.mjs',
+  task:     'task.mjs',
   share:    'share.mjs',
   'generate-pat': 'generate-pat.mjs',
   'hash-password': 'hash-password.mjs',
@@ -77,6 +78,8 @@ const COMMANDS = {
   brain:    'brain.mjs',
   config:   'config.mjs',
   skill:    'skill.mjs',
+  secret:   'secret.mjs',
+  secrets:  'secret.mjs',
   help:     'help.mjs',
   collab:   'collab.mjs',
   command:  'command.mjs',
@@ -84,70 +87,42 @@ const COMMANDS = {
 };
 function printHelp() {
   console.log(`
-  total-recall — Sovereign AI System CLI
+  total-recall — Portable personal memory for any IDE
 
   Usage: total-recall <command> [options]
 
-  Commands:
-    init                Bootstrap Total Recall into an existing project repo
-    setup               Interactive wizard: provider → API key → provision → connect IDEs
-    start [--port N]    Start the brain server in the foreground (alias for running the server directly)
-    deploy              Provision a target machine (Caddy/TLS, system services, VFS setup, environment checks)
-    doctor              Run environment diagnostics and verify dependency health
-    dream               Manually trigger a dream cycle (Light → REM → Deep)
-    lint                Validate all vault nodes against SSSS schema v2
-    daemon <start|stop|status>  Manage the background daemon
-    status              Show server health, active deploy mode, dashboard URL, tunnel PID, and daemon state
-    backup [options]    Create encrypted VFS tarball, push git backup, or Obsidian sync
-    restore <path>      Restore from an encrypted backup
-    upgrade --model <n> Swap the kernel model (e.g., gemma5-32b)
-    friction            Analyze logs to detect workflow bottlenecks
-    chat                Interactive terminal chat with the Sovereign OS
-    remember            Remember new rules, preferences, or corrections in real time
-    forget <slug>       Delete a memory node by slug (with auto-recompile)
-    recall              Recall semantic memory or search existing knowledge/rules/session-history
-    map                 Visualize memory vault category trees and link relationships
-    brain <cmd>         Manage registered brains and context resolution (list, status, register, unregister)
-    research <command>  Manage/query ongoing autonomous research agenda (list, add, show, report, cancel)
-    share [url]         Share a URL or text snippet to the brain (auto-routes to remember or research)
-    config <get|set>    Read, write, and toggle UI, security, and budget settings
-    skill <command>     Manage AI agent skills: find, install, scan, list, remove
-    generate-pat        Issue a PAT; stores only a hash in keys.jsonl
-    hash-password       Generate a bcrypt dashboard password hash
-    reset-password      Reset the dashboard admin password and force a change on login
-    compile             Alias for rebuild: deterministically rebuild projections
-    rebuild [--check]   Detect index drift and deterministically rebuild projections
-    snapshot            Manage fast point-in-time VFS snapshots and rollbacks
-    migrate             Run backwards-incompatible SSSS schema migrations
-    import [--dir <path>] [--force]   Import existing AGENTS.md, .cursorrules, CLAUDE.md, etc.
-                                     into the vault (run before compile on first install)
-    ingest [--watch]    Ingest IDE conversation logs (Claude Code, Codex, Gemini, etc.)
-    ingest google-takeout <path>  Ingest Google Takeout export data
-    export <path> --okf Export memory vault as an OKF-compliant bundle
-    connect <client>    Configure Cursor, Claude Code, Codex, UltraChat, Gemini, etc.
-    sync [--watch]      Pull remote brain instructions into the current workspace
-    relay <cmd>         Local background relay: ship IDE sessions to remote brain
-                        Commands: start | stop | status | once | install | uninstall
-    uninstall           Completely uninstall all services, launchd agents, VFS, and shims
-    help <topic>        Interactive offline documentation, SSSS reference, and system help
-    collab              Start Express/WebSockets backend & React collab sandbox dev server
-    command <create|read|update|remove|list>  Manage custom project-local CLI commands
+  Core (default product path):
+    init [--project]    Bootstrap global or project brain + openwiki
+    connect <client>    Wire IDE / Obsidian / http-api host
+    remember / forget   Write-path memory (SSSS vault)
+    recall              Read-path hybrid search
+    compile             Rebuild instruction surfaces
+    dream               Memory consolidation cycle
+    task <cmd>          Enqueue daemon work (open envelope)
+    daemon <cmd>        Background worker (start|stop|status)
+    skill <cmd>         Skills registry, deploy, multi-repo sync
+    secret <cmd>        Secrets store + usage (not the vault)
+    brain <cmd>         Register / ensure any project brain
+    status / doctor     Health and diagnostics
+
+  Optional:
+    research, share, ingest, import, export, relay, setup, deploy,
+    start, backup, restore, config, map, generate-pat, lint, sync, …
+
+  Full inventory: docs/reference/CLI_INVENTORY.md
+  Default story: init → connect → remember/recall → dream (no LLM required)
 
-
-
-
-  Autonomous operations (sync, compile, backup) are now handled by the
-  Cloud Agent via SSSS task nodes in .agent/scheduler/queue/.
-  The agent cron trigger (every 5 min) processes these automatically.
+  Examples:
+    npx total-recall init --project
+    npx total-recall connect claude-code
+    npx total-recall remember fact "The API uses port 3000."
+    npx total-recall dream
+    npx total-recall skill track .
+    npx total-recall secret list
 
   Options:
-    --help, -h          Show this help message
+    --help, -h          Show this help
     --version, -v       Show version
-
-  Examples:
-    npx total-recall deploy
-    npx total-recall dream
-    npx total-recall daemon status
 `);
 }
 

package/extension/background.js

@@ -40,7 +40,7 @@ chrome.contextMenus.onClicked.addListener(async (info, tab) => {
       await share({
         url: tab.url,
         title: tab.title,
-        action: 'research',
+        action: 'remember',
         source: 'chrome-extension'
       });
       showBadge('\u2713', '#22c55e');

package/extension/options/options.js

@@ -17,7 +17,8 @@
       chrome.storage.local.get(['brainUrl', 'pat']),
       chrome.storage.sync.get([
       'blocklist',
-      'captureGranularity'
+      'captureGranularity',
+      'passiveTracking'
       ])
     ]);
     const settings = { ...syncSettings, ...localSettings };
@@ -32,7 +33,10 @@
     }
 
     // Granularity radio
-    const granularity = settings.captureGranularity || 'minimal';
+    let granularity = settings.captureGranularity || 'minimal';
+    if (settings.passiveTracking === false) {
+      granularity = 'off';
+    }
     const radio = document.querySelector(`input[name="granularity"][value="${granularity}"]`);
     if (radio) radio.checked = true;
   }
@@ -49,6 +53,7 @@
       // Get selected granularity
       const granularityRadio = document.querySelector('input[name="granularity"]:checked');
       const captureGranularity = granularityRadio ? granularityRadio.value : 'minimal';
+      const passiveTracking = captureGranularity !== 'off';
 
       await chrome.storage.local.set({
         brainUrl: brainUrlInput.value.trim() || 'http://127.0.0.1:3000',
@@ -57,7 +62,8 @@
       await chrome.storage.sync.remove('pat');
       await chrome.storage.sync.set({
         blocklist,
-        captureGranularity
+        captureGranularity,
+        passiveTracking
       });
 
       showSaveStatus('Settings saved!', false);

package/extension/sidepanel/sidepanel.js

@@ -558,8 +558,6 @@ Provide helpful analysis grounded on both the user's brain memory and this page
   settingsTrackingToggle.addEventListener('change', async () => {
     const passiveTracking = settingsTrackingToggle.checked;
     await chrome.storage.sync.set({ passiveTracking });
-    // Signal active preference back to service worker
-    chrome.runtime.sendMessage({ type: 'UPDATE_PREFERENCES', passiveTracking });
     showToast(`Passive tracking ${passiveTracking ? 'enabled' : 'disabled'}`);
   });
 

package/frontend/index.html

@@ -3,13 +3,14 @@
   <head>
     <meta charset="UTF-8" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-    <meta name="description" content="Total Recall — Sovereign AI Brain Dashboard. Memory explorer, chat interface, code sandbox, and system health monitoring." />
-    <meta name="theme-color" content="#0a0a0f" />
+    <meta name="description" content="Total Recall — portable personal memory for any IDE. Memory explorer, chat, dream consolidation, and system health." />
+    <meta name="theme-color" content="#070b14" />
+    <link rel="icon" type="image/svg+xml" href="/favicon.svg" />
     <link rel="manifest" href="/manifest.json" />
     <link rel="preconnect" href="https://fonts.googleapis.com" />
     <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
     <link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet" />
-    <title>Total Recall — Sovereign Brain</title>
+    <title>Total Recall — Portable Memory</title>
   </head>
   <body>
     <div id="root"></div>