mixdog 1.0.12 → 1.1.0

package/README.md

@@ -1,199 +1,80 @@
 <p align="center">
-  <img src="https://raw.githubusercontent.com/tribgames/mixdog/main/apps/desktop/build/mixdog.png" alt="" width="96">
+  <img src="https://raw.githubusercontent.com/tribgames/mixdog/main/apps/desktop/build/mixdog.png" width="80" alt="Mixdog">
 </p>
 
 <h1 align="center">Mixdog</h1>
 
 <p align="center">
-  <b>Same model. Same performance. 63% fewer tokens.</b><br>
-  Free, open-source desktop coding agent for Windows.
+  The open-source coding agent that uses 63% fewer tokens.
 </p>
 
 <p align="center">
-  <a href="https://github.com/tribgames/mixdog/releases/latest/download/mixdog-desktop-win-x64.exe">
-    <img src="https://raw.githubusercontent.com/tribgames/mixdog/main/docs/assets/download-windows.svg" alt="Download Mixdog for Windows x64" width="320">
-  </a>
+  <a href="https://github.com/tribgames/mixdog/releases/latest/download/mixdog-desktop-win-x64.exe"><img src="https://raw.githubusercontent.com/tribgames/mixdog/main/docs/assets/download-windows.svg" alt="Download for Windows" height="48"></a>
+  &nbsp;
+  <a href="https://github.com/tribgames/mixdog/releases/latest/download/mixdog-desktop-mac-arm64.dmg"><img src="https://raw.githubusercontent.com/tribgames/mixdog/main/docs/assets/download-macos.svg" alt="Download for macOS" height="48"></a>
+  <br>
+  <sub>macOS button is for Apple Silicon (M1 and later). Intel Mac? <a href="https://github.com/tribgames/mixdog/releases/latest/download/mixdog-desktop-mac-x64.dmg">Download for Intel Mac</a></sub>
 </p>
 
 <p align="center">
-  <a href="https://github.com/tribgames/mixdog/releases/latest"><img src="https://img.shields.io/github/v/release/tribgames/mixdog?label=release" alt="Latest release"></a>
-  <a href="LICENSE"><img src="https://img.shields.io/badge/license-Apache--2.0-blue" alt="Apache-2.0 license"></a>
-  <img src="https://img.shields.io/badge/platform-Windows%20x64-0078D4" alt="Windows x64">
+  <a href="https://github.com/tribgames/mixdog/releases/latest"><img alt="Release" src="https://img.shields.io/github/v/release/tribgames/mixdog?style=flat-square"></a>
+  <a href="https://www.npmjs.com/package/mixdog"><img alt="npm" src="https://img.shields.io/npm/v/mixdog?style=flat-square"></a>
+  <a href="LICENSE"><img alt="License" src="https://img.shields.io/badge/license-Apache--2.0-blue?style=flat-square"></a>
 </p>
 
 <p align="center">
-  <sub>The installer is not code-signed yet. If Windows SmartScreen appears,
-  choose <b>More info → Run anyway</b>.</sub>
+  <a href="https://aiscroll.io/mixdog/"><b>Website</b></a> ·
+  <a href="benchmarks/terminal-bench-2.1/"><b>Benchmarks</b></a> ·
+  <a href="#documentation"><b>Docs</b></a>
 </p>
 
-<p align="center">
-  <img src="https://raw.githubusercontent.com/tribgames/mixdog/main/docs/assets/desktop.png" alt="Mixdog Desktop" width="860">
-</p>
-
-<a name="benchmarks"></a>
-## Same performance, 63% fewer tokens
-
-Mixdog and Codex CLI ran Terminal-Bench 2.1 on the same model, GPT-5.6 Sol
-xhigh. Both passed the same share of tasks. Mixdog used 63% fewer tokens.
+https://github.com/user-attachments/assets/af515dc9-b0fb-4b6a-aaa1-1d82db792ed0
 
-| | Mixdog | Codex CLI | |
-| --- | --- | --- | --- |
-| **Total tokens** (incl. cached input) | **156.5M** | 421.4M | **63% fewer** |
-| Success rate | **86.5%** (385/445) | 86.1% (383/445) | matched |
-| Priced cost per trial | **$0.476** | $0.782 | 39% lower |
-| Median first request | **4.7k** | 14.4k | 67% smaller |
-| Median final context | **18.5k** | 34.3k | 46% smaller |
-| Wall time per trial | **415s** | 437s | matched |
+## Highlights
 
-![Terminal-Bench 2.1: Mixdog with GPT-5.6 Sol xhigh versus Codex CLI](https://raw.githubusercontent.com/tribgames/mixdog/main/benchmarks/terminal-bench-2.1/tb21-sol-vs-codex.svg)
+- [63% fewer tokens](#benchmarks) than Codex CLI on the same model, at the
+  same success rate.
+- [**Auto reasoning effort**](docs/auto-reasoning-effort.md) — each message
+  and tool step gets the effort it needs, judged in about 0.1 s by a 130 MB
+  model that runs locally on the CPU. On by default.
+- Use the **Claude, ChatGPT, or Grok** subscription you already have, API
+  keys, or local models on your own GPU.
+- Mix models by role and scale from **Solo to Swarm** — a strong model leads
+  while cheaper ones search, edit, and review in parallel.
+- A full workspace: editor, terminals, Git and GitHub, code graph, and
+  language servers.
+- Keeps working on its own with **Goals, schedules, and webhooks** — and
+  continues from your phone, end-to-end encrypted.
+- Goes beyond code with **browser and desktop control**, Office and PDF
+  documents, and image and video generation.
+- Runs on Windows and macOS. No config files.
 
-- **The official leaderboard protocol.** All 89 tasks, five trials each — 445
-  trials per side — with unmodified task timeouts and resources, scored by the
-  official Harbor verifier.
-- **Same conditions on both sides.** Same model and reasoning level, the same
-  kind of subscription sign-in, fast mode off, no retries of task failures or
-  agent timeouts.
-- **Everything is published.** Verdicts, verifier output, and usage snapshots
-  for both sides, plus the scripts that recompute every number, are in
-  [`benchmarks/terminal-bench-2.1/`](benchmarks/terminal-bench-2.1/).
+## Installation
 
-### Across the 89 tasks
+Download the desktop app:
 
-Mixdog used fewer tokens on 87 of the 89 tasks. The median task used 68% fewer.
-
-| Token change | Tasks |
+| Platform | Download |
 | --- | --- |
-| 75% fewer or better | 24 |
-| 50–75% fewer | 42 |
-| 25–50% fewer | 19 |
-| 0–25% fewer | 2 |
-| More tokens | 2 |
-
-Pass counts were equal on 60 tasks; Mixdog passed more trials on 14 and Codex
-CLI on 15. Mixdog used more tokens on `mteb-leaderboard` (6.8×) and
-`crack-7z-hash` (1.4×). Every task is listed in
-[`results.md`](benchmarks/terminal-bench-2.1/results.md).
-
-<details>
-<summary><b>Claude Opus 5 — Mixdog vs Claude Code</b> (single pass, 89 trials each)</summary>
+| Windows (x64) | [`mixdog-desktop-win-x64.exe`](https://github.com/tribgames/mixdog/releases/latest/download/mixdog-desktop-win-x64.exe) |
+| macOS (Apple Silicon) | [`mixdog-desktop-mac-arm64.dmg`](https://github.com/tribgames/mixdog/releases/latest/download/mixdog-desktop-mac-arm64.dmg) |
+| macOS (Intel) | [`mixdog-desktop-mac-x64.dmg`](https://github.com/tribgames/mixdog/releases/latest/download/mixdog-desktop-mac-x64.dmg) |
 
-| | Mixdog | Claude Code | |
-| --- | --- | --- | --- |
-| Solved | **79/89** | 77/89 | +2 tasks |
-| Priced cost per run | **$104.29** | $129.21 | 19% lower |
-| Median final context | **27.6k** | 38.2k | 28% smaller |
-| Wall time per trial | **610s** | 708s | 1.16× faster |
+Then sign in with a ChatGPT, Claude, or Grok account, or paste an API key. A
+five-step tutorial walks you through the rest.
 
-![Terminal-Bench 2.1: Mixdog with Claude Opus 5 versus Claude Code](https://raw.githubusercontent.com/tribgames/mixdog/main/benchmarks/terminal-bench-2.1/tb21-opus-vs-claude-code.svg)
+> [!NOTE]
+> The Windows installer is not code-signed yet. If SmartScreen appears,
+> choose **More info → Run anyway**. The Mac app is signed and notarized.
 
-One trial per task, so per-task differences sit within run-to-run variance.
+Or run it in the terminal (Node.js 22.19+ or 24+):
 
-</details>
-
-<sub>Codex CLI 0.151.0. Mixdog ran single-model, single-session — no
-sub-agents or helper models. Cost values both sides at the same API list
-rates, not subscription charges. Results measure the pinned source revision.
-The leaderboard is not accepting community submissions, so the runs are
-published here instead.</sub>
-
-## How it uses fewer tokens
-
-- **A light start.** System instructions and tool descriptions are kept to
-  what the model needs.
-  *67% smaller first request — 4.7k tokens against 14.4k.*
-- **Fewer round trips.** The agent picks the right tool first and runs
-  independent actions in one batch.
-  *33% fewer model requests — a median of 10 per trial against 16.*
-- **Smaller requests.** Tools return only the part that was asked for, and
-  95% of terminal output is filtered before it reaches the model.
-  *46% less input per request — 24.3k tokens against 44.9k.*
-- **Shorter output.** Less filler and repetition in replies.
-  *8% fewer output tokens.*
-
-The benchmark ran one model in one session. In everyday use, more applies on
-top of that:
-
-- **Auto-clear.** After a long break, the conversation is compacted before
-  work resumes.
-- **Light compaction.** Long sessions continue from a structured handoff
-  instead of the full history.
-- **Database memory.** Past work is stored and searched, so the prompt does
-  not grow with the archive.
-- **Code graph.** Query a project's structure instead of reading whole files.
-- **Orchestration.** Hand routine work to cheaper models and keep the
-  expensive one for the steps that need it.
-
-See [Context efficiency](docs/context-efficiency.md) for how each layer works.
-
-## A model for each job
-
-Not every step needs your most expensive model.
-
-- **Mix models by role.** Let a strong model lead and plan while cheaper ones
-  search, edit, and review.
-- **Dial delegation from Solo to Swarm.** The lead can do the work itself or
-  coordinate a team of agents working in parallel.
-- **Workflows you can read.** Workflows and agents are plain Markdown
-  (`WORKFLOW.md`, `AGENT.md`) with visual editors in the app.
-- **Every account in one place.** Claude, ChatGPT, and Grok subscriptions, API
-  keys, and local models side by side.
-- **Run models on your own GPU.** The built-in Local Provider downloads and
-  runs models inside the app.
-- **Know what it costs.** Usage by provider and model — input, output, cache
-  hits, and cost — plus quota windows and resets.
-
-## Everything in one window
-
-- **A full workspace.** Tabs and split panes, Monaco editor, terminals, file
-  explorer, and language servers that start when you open a file.
-- **Git and GitHub built in.** Review changes, commit, and handle pull
-  requests, issues, Actions, and releases without leaving the app.
-- **Code that stays clean.** The code graph maps a project's structure, and
-  Code Tidy formats, lints, and applies structural fixes.
-- **Memory that lasts.** Mixdog remembers what matters across sessions and
-  searches past work on demand.
-- **Extend it.** Add skills, MCP servers, and plugins from one Extensions view.
-
-## It keeps working when you step away
-
-- **Goals.** Set a Goal and the session keeps working toward it — pause and
-  resume whenever you like.
-- **Schedules and webhooks.** Run a task on a timer or whenever a URL is
-  called.
-- **Your phone is the remote.** Scan a QR code to continue the same live
-  session from your phone, end-to-end encrypted, with a notification when the
-  task finishes.
-
-## Beyond code
-
-- **Browser Use.** Operate real, signed-in web pages, with Chrome profile
-  import.
-- **Computer Use.** Operate native Windows apps with guarded input and an
-  on-screen Stop control.
-- **Documents.** Create and edit Word, Excel, PowerPoint, and PDF with
-  rendered previews.
-- **Image and video Studio.** Generate and edit with a local gallery.
-
-Browser Use and Computer Use are opt-in and ask for approval before their
-first live call in each session.
-
-## Get started
-
-[Download the installer](https://github.com/tribgames/mixdog/releases/latest/download/mixdog-desktop-win-x64.exe)
-and sign in. No config files, no YAML, no terminal.
-
-- **One sign-in and you are working.** Use the ChatGPT, Claude, or Grok
-  subscription you already pay for, or paste an API key. Mixdog picks the
-  model for you.
-- **Advanced setups, one switch each.** A team of agents, a model per role,
-  long-term memory, phone access — each is a toggle, not a config file.
-- **You won't get lost.** A short tutorial gets you set up in five quick
-  steps.
-- **Batteries included.** Git, Memory, Browser, and Office tools ship with the
-  app. Turn on what you need.
+```bash
+npm install -g mixdog
+mixdog
+```
 
 <details>
-<summary><b>Supported providers</b></summary>
+<summary>Supported providers</summary>
 
 - Anthropic API keys and Claude account OAuth
 - OpenAI API keys and ChatGPT/Codex account OAuth
@@ -209,14 +90,7 @@ Cursor and Antigravity (Gemini) OAuth are off by default under
 </details>
 
 <details>
-<summary><b>Command line</b></summary>
-
-Mixdog also runs in the terminal. Requires Node.js 22.19+ (22.x) or 24+.
-
-```bash
-npm install -g mixdog
-mixdog
-```
+<summary>Headless <code>mixdog exec</code></summary>
 
 `mixdog exec` runs one non-interactive, single-model session without personal
 memory, prior sessions, skills, MCP servers, or plugins:
@@ -226,15 +100,35 @@ mixdog exec --provider openai-oauth --model gpt-5.6-sol --effort xhigh "fix the
 ```
 
 Web search is off by default (`--web-search` enables it). This does not block
-shell networking — headless exec is not an offline sandbox.
-
-Run `mixdog --help` for the full reference.
+shell networking — headless exec is not an offline sandbox. Run
+`mixdog --help` for the full reference.
 
 </details>
 
-## Docs
+## Benchmarks
+
+<p align="center">
+  <picture>
+    <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/tribgames/mixdog/main/benchmarks/terminal-bench-2.1/tb21-sol-vs-codex-dark.svg">
+    <source media="(prefers-color-scheme: light)" srcset="https://raw.githubusercontent.com/tribgames/mixdog/main/benchmarks/terminal-bench-2.1/tb21-sol-vs-codex-light.svg">
+    <img alt="Bar chart: Mixdog vs Codex CLI on Terminal-Bench 2.1" src="https://raw.githubusercontent.com/tribgames/mixdog/main/benchmarks/terminal-bench-2.1/tb21-sol-vs-codex-light.svg">
+  </picture>
+</p>
+
+<p align="center">
+  <i>All 89 Terminal-Bench 2.1 tasks × 5 trials on GPT-5.6 Sol xhigh, scored by the official Harbor verifier.</i>
+</p>
+
+Mixdog used fewer tokens on 87 of 89 tasks. With Claude Opus 5, it solved 2
+more tasks than Claude Code at 19% lower cost. Every verdict, log, and script
+is published in [`benchmarks/terminal-bench-2.1/`](benchmarks/terminal-bench-2.1/),
+and [Context efficiency](docs/context-efficiency.md) explains where the
+savings come from.
+
+## Documentation
 
 - [Context efficiency](docs/context-efficiency.md)
+- [Auto reasoning effort](docs/auto-reasoning-effort.md)
 - [Benchmarks](benchmarks/terminal-bench-2.1/)
 - [Code Tidy](docs/code-tidy.md)
 - [Git & GitHub](docs/git-github-integration.md)
@@ -250,5 +144,5 @@ you tokens, a star helps others find it.
 
 ## License
 
-Mixdog is licensed under [Apache-2.0](LICENSE).
-Third-party components retain their respective licenses; see [NOTICE.md](NOTICE.md).
+Mixdog is licensed under [Apache-2.0](LICENSE). Third-party components retain
+their respective licenses; see [NOTICE.md](NOTICE.md).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mixdog",
-  "version": "1.0.12",
+  "version": "1.1.0",
   "private": false,
   "type": "module",
   "description": "Open-source coding agent that matches Codex on Terminal-Bench 2.1 with 63% fewer tokens.",
@@ -57,7 +57,9 @@
   ],
   "scripts": {
     "postinstall": "node scripts/prune-embedding-runtime.mjs && node scripts/prepare-native-assets.mjs",
-    "prepack": "npm run build:tui",
+    "prepack": "npm run build:shipped-history && npm run build:tui",
+    "build:shipped-history": "node scripts/build-shipped-history.mjs",
+    "check:shipped-history": "node scripts/build-shipped-history.mjs --check",
     "prepublishOnly": "node -e \"if(!process.env.CI){console.error('local npm publish is disabled — run npm run release:patch');process.exit(1)}\"",
     "start": "node src/cli.mjs",
     "test": "node scripts/test.mjs",

package/scripts/build-shipped-history.mjs

@@ -0,0 +1,131 @@
+#!/usr/bin/env node
+/**
+ * scripts/build-shipped-history.mjs — emit src/defaults/shipped-history.json:
+ * normalised-content hashes of every past shipped version of the agents,
+ * workflows, output styles and bundled skills, read from git history.
+ *
+ * The startup migration (src/session-runtime/services/seeds.mjs) treats a data
+ * dir copy matching any of these hashes as "unchanged" and moves it to a
+ * backup so the current shipped version applies. Hash and normalisation code
+ * is shared with the runtime (src/runtime/shared/shipped-definitions.mjs).
+ *
+ * Run:  node scripts/build-shipped-history.mjs   (also part of `prepack`)
+ *       node scripts/build-shipped-history.mjs --check   (release workflow)
+ * A shallow clone has no history: the committed file is left untouched.
+ */
+import { execFileSync } from 'node:child_process';
+import { readFileSync, writeFileSync } from 'node:fs';
+import { dirname, join } from 'node:path';
+import { fileURLToPath } from 'node:url';
+import {
+  SHIPPED_HISTORY_FILE,
+  hashDefinition,
+  hashText,
+  hashTree,
+  packDefinitionFrom,
+} from '../src/runtime/shared/shipped-definitions.mjs';
+
+const ROOT = join(dirname(fileURLToPath(import.meta.url)), '..');
+const OUT = join(ROOT, 'src', ...SHIPPED_HISTORY_FILE);
+// --check: write nothing, exit 1 when the committed file differs (release gate).
+const CHECK = process.argv.includes('--check');
+
+const git = (...args) =>
+  execFileSync('git', args, { cwd: ROOT, maxBuffer: 1 << 28, stdio: ['ignore', 'pipe', 'inherit'] });
+const lines = (buffer) => buffer.toString('utf8').split('\n').filter(Boolean);
+
+if (git('rev-parse', '--is-shallow-repository').toString().trim() === 'true') {
+  console.warn('build-shipped-history: shallow clone, keeping the committed src/defaults/shipped-history.json');
+  process.exit(0);
+}
+
+const blobs = new Map();
+const blob = (sha) => {
+  if (!blobs.has(sha)) blobs.set(sha, git('cat-file', 'blob', sha));
+  return blobs.get(sha);
+};
+
+/** Names (path segment after `prefix`) that ever existed under prefix. */
+function everNames(prefix) {
+  const names = new Set();
+  for (const file of lines(git('log', '--name-only', '--format=', '--', prefix))) {
+    if (file.startsWith(`${prefix}/`)) names.add(file.slice(prefix.length + 1).split('/')[0]);
+  }
+  return [...names].sort();
+}
+
+/** Every version of `path`: Map of relative path -> blob sha, one per commit that touched it. */
+function versionsOf(path) {
+  const versions = [];
+  for (const commit of lines(git('log', '--format=%H', '--', path))) {
+    const files = new Map();
+    for (const row of lines(git('ls-tree', '-r', commit, '--', path))) {
+      const [meta, file] = row.split('\t');
+      files.set(file.slice(path.length + 1) || file.split('/').pop(), meta.split(' ')[2]);
+    }
+    if (files.size) versions.push(files);
+  }
+  return versions;
+}
+
+const unique = (hashes) => [...new Set(hashes)].sort();
+
+function packHistory(dir, entry) {
+  const out = {};
+  for (const id of everNames(dir)) {
+    const hashes = [];
+    for (const files of versionsOf(`${dir}/${id}`)) {
+      const read = (name) => (files.has(name) ? blob(files.get(name)).toString('utf8') : null);
+      const definition = packDefinitionFrom(read, entry, id);
+      if (definition) hashes.push(hashDefinition(definition));
+    }
+    if (hashes.length) out[id] = unique(hashes);
+  }
+  return out;
+}
+
+function outputStyleHistory() {
+  const out = {};
+  for (const name of everNames('src/output-styles')) {
+    const hashes = versionsOf(`src/output-styles/${name}`).map((files) => hashText(blob([...files.values()][0])));
+    if (hashes.length) out[name] = unique(hashes);
+  }
+  return out;
+}
+
+function skillHistory() {
+  const out = {};
+  for (const name of everNames('src/defaults/skills')) {
+    const hashes = versionsOf(`src/defaults/skills/${name}`).map((files) =>
+      hashTree([...files].map(([rel, sha]) => [rel, blob(sha)]))
+    );
+    if (hashes.length) out[name] = unique(hashes);
+  }
+  return out;
+}
+
+const history = {
+  agents: packHistory('src/agents', 'AGENT.md'),
+  workflows: packHistory('src/workflows', 'WORKFLOW.md'),
+  outputStyles: outputStyleHistory(),
+  skills: skillHistory(),
+};
+const text = `${JSON.stringify(history, null, 2)}\n`;
+if (CHECK) {
+  let committed = '';
+  try {
+    committed = readFileSync(OUT, 'utf8');
+  } catch {
+    committed = '';
+  }
+  if (committed !== text) {
+    console.error(
+      'build-shipped-history: src/defaults/shipped-history.json is stale; run `npm run build:shipped-history` and commit it'
+    );
+    process.exit(1);
+  }
+  console.log('build-shipped-history: committed history is current');
+} else {
+  writeFileSync(OUT, text);
+  console.log(`build-shipped-history: wrote ${OUT}`);
+}