symbiot 0.28.2 → 0.28.4

package/README.md

@@ -41,49 +41,42 @@ no install — it's the same CLI. Press Ctrl+C (or click Quit) to stop.
 
 Tabs:
 
-- **Map** — a live, interactive node graph of your work, built from your local git:
-  you at the centre, your repos, and the languages and tools they share (so related
-  projects cluster), plus the coding **agents and editors** you have installed and
-  the **AI** currently powering Symbiot. Project folders that **aren't git repos**
-  show up too (see [Where it looks](#where-it-looks)). **Scroll to zoom, drag to
-  pan, click a node.** Clicking a repo shows its branch/commits/uncommitted/stack, a
+- **Map** (the landing view) — a live node graph of your work from your local git:
+  you at the centre, your repos, the languages and tools they share (so related
+  projects cluster), the coding **agents and editors** you have installed, and the
+  **AI** powering Symbiot. Project folders that **aren't git repos** show up too
+  (see [Where it looks](#where-it-looks)). **Scroll to zoom, drag to pan, click a
+  node.** A repo's panel shows its branch, commits, uncommitted work and stack, a
   **Suggest next steps** button, and an **AI review**: what the project does, a
-  one-line verdict on whether it needs new work at all (it's told to prefer
-  stabilising over piling on features), and a short list of ideas you can tick
-  straight into **Tasks**. The graph itself **needs no AI key** (pure local data);
-  reviews and suggestions use your chosen model (free with local Ollama). While
-  the scan runs, the Map shows live progress (which phase, how many repos, which
-  one). This is the landing view.
+  one-line verdict on whether it needs new work at all (it prefers stabilising over
+  new features), and ideas you can tick straight into **Tasks**. The graph needs
+  **no AI key**; reviews and suggestions use your chosen model. While the scan runs
+  the Map shows its progress.
 - **Drift** — the [`symbiot drift`](#whats-out-of-sync-symbiot-drift) report, with
   "fetch latest" and "check CI" toggles.
 - **Week / Standup / Todo** — the write-ups (these use your chosen AI).
 - **Tasks** — a checklist, grouped by kind (Fixes, Tests & CI, Docs, …). Filter by
   type or repo, then **Send to repos** to hand just those to your agent (see
   [`symbiot push`](#hand-tasks-to-your-coding-agent-symbiot-push)). What your agent
-  finishes lands in **Awaiting your review**, where you approve or send it back (see
-  [Review and approve](#review-and-approve-the-agents-work)). Ticking a task yourself
-  marks it done and it **auto-archives**. An **archived** view lets you restore any
-  of them.
-- **Agents** — every agent run Symbiot has started (and local-model downloads),
-  with live status, elapsed time, exit code and the tail of its output.
-- **Settings** — pick your AI, the folders to scan, your agent command, and get
-  model recommendations or a one-click local model.
+  finishes lands in **Awaiting your review** (see
+  [Review and approve](#review-and-approve-the-agents-work)). A task you tick
+  yourself is done and **auto-archives**; the archived view can restore it.
+- **Agents** — every agent run Symbiot has started (and local-model downloads):
+  live status, elapsed time, exit code, the tail of its output, and **what it
+  did** — files changed and commits made, read from git, whichever agent it was.
+- **Settings** — your AI, the folders to scan, your agent command, model
+  recommendations and a one-click local model.
 
 **Stays current by itself.** The app checks npm for a newer Symbiot every couple of
-minutes (and whenever you come back to the window). When there is one, a banner
-offers **Update & restart**: it runs `npm install -g symbiot@latest`, relaunches the
-app on the same address, and the open window reloads itself onto the new version —
-no terminal, no reopening. The version you're on is shown next to the name. The
-same mechanism means that if you restart `symbiot app` yourself, the open window
-reconnects and reloads on its own.
+minutes (and whenever you come back to the window); the version you're on is shown
+next to the name. When there's a newer one, **Update & restart** installs that exact
+version, relaunches the app on the same address, and the open window reloads itself
+onto it. Restart `symbiot app` yourself and the open window reconnects the same way.
 
 The app listens on port **7391** so its address survives restarts; set
 `SYMBIOT_PORT` to use another (it falls back to a free port if that one is busy).
-
-**Cross-platform:** the app works on **Linux, macOS, and Windows** — the server
-and UI are just a local web page. The chrome-less window is detected per-OS
-(PATH on Linux, the `/Applications` bundle on macOS, `Program Files` on Windows);
-where no Chromium-family browser is found it opens your default browser instead.
+It works on **Linux, macOS and Windows**; with no Chromium-family browser to open
+the chrome-less window, it uses your default browser.
 
 ## Hand tasks to your coding agent: `symbiot push`
 
@@ -114,7 +107,8 @@ size of the uncommitted change and a **Show diff** button. For each repo:
   on a feature branch it commits and opens the PR from there. Each step that can't
   happen stops there and says why: no `origin` remote means a local commit only, and
   without `gh` it pushes and stops. Your commit is never lost. Then the tasks are
-  archived with their commit and PR link.
+  archived with their commit and PR link. With no changes to commit the button
+  reads **Approve → archive** and just archives them.
 - **↩ (send back)** is for one that isn't right: it reopens the task and unticks it in
   `TASKS.md`, so the next **Send to repos** hands it to the agent again.
 
@@ -133,9 +127,10 @@ aider --message "{prompt}" --yes
 code {dir}
 ```
 
-Settings shows **one-click presets** for what's installed on your machine:
+Settings shows **one-click presets** for what's installed on your machine. A
+preset only fills in the command box; the saved command is what runs.
 
-- **Agents:** Claude Code (*make changes*, or *plan only* which asks first), Codex
+- **Agents** (all make changes, ready for your review): Claude Code, Codex
   (OpenAI), Aider, Cursor agent, Gemini CLI.
 - **Orca IDE** (any OS): opens the repo in Orca — either just the repo, to use
   Orca's own agent, or with Claude running in a new tab. It launches Orca if it's
@@ -146,7 +141,7 @@ Settings shows **one-click presets** for what's installed on your machine:
 Each run is logged to `.symbiot/agent.log` in the repo and shown live in the
 **Agents** tab. The command is saved as `agentCmd` in
 `~/.config/symbiot/config.json` — it's your command, Symbiot only fills in
-`{dir}` and `{prompt}`.
+`{dir}` and `{prompt}`. A command saved from an older preset keeps working as-is.
 
 ## What's out of sync? `symbiot drift`
 
@@ -300,16 +295,26 @@ conventions every repo review must respect, alongside each repo's own
 npm test
 ```
 
-Runs three suites, all against throwaway repos and an isolated `HOME`:
+Runs four suites, all against throwaway repos and an isolated `HOME`:
 
+- `test/load.mjs` — the shipped files parse and load: `node --check` on each
+  module, `ui.mjs` imported on its own with its page's JavaScript parsed, every
+  local import listed in package.json `"files"`, and the `bin` entry point shipped.
 - `test/fixtures.mjs` — accuracy fixtures: the facts Symbiot collects (identity
   matching, stale checkouts, worktrees, drift) and the `symbiot drift` report as
-  printed, plus the scan time limit.
+  printed, the scan time limit, the agent handoff, and the review → approve cycle.
 - `test/smoke.mjs` — boots `symbiot app`, runs the page's own JavaScript against
-  a fake DOM (every tab, every button), and hits every endpoint the UI calls.
+  a fake DOM (every tab, every button), hits every endpoint the UI calls with the
+  method the UI uses, and checks POST-only endpoints refuse GET and `/api` needs
+  the token.
 - `test/install.mjs` — runs the CLI through a bin symlink, then `npm pack` +
-  global install into a temp prefix and runs `symbiot help`. Needs npm registry
-  access for dependencies; `SYMBIOT_SKIP_INSTALL_TEST=1` skips that part.
+  global install into a temp prefix: `symbiot help`, every shipped file present,
+  and the installed `symbiot app` serving its page. Needs npm registry access for
+  dependencies; `SYMBIOT_SKIP_INSTALL_TEST=1` skips that part.
+
+`SYMBIOT_NO_OPEN=1` stops `symbiot app` opening a window (the tests set it).
+Releases publish from a version tag via `.github/workflows/publish.yml`, and only
+after the full suite passes in CI.
 
 ---
 

package/index.mjs

@@ -182,7 +182,12 @@ function shipChanges(path, texts, opts = {}) {
     branch = stem; for (let i = 2; git(path, ["rev-parse", "--verify", "-q", "refs/heads/" + branch]).ok; i++) branch = `${stem}-${i}`;
     const sw = git(path, ["switch", "-c", branch]); if (!sw.ok) return { error: "Could not create branch: " + sw.err };
   }
-  const add = git(path, ["add", "-A", ...NOT_SYMBIOT]); if (!add.ok) return { error: "git add failed: " + add.err, branch };
+  // Stage everything, then drop .symbiot. A `. :(exclude).symbiot` pathspec
+  // warns+exits-1 once .symbiot is gitignored ("paths are ignored, use -f"),
+  // which falsely aborted the ship. `add -A` skips gitignored paths silently;
+  // the reset also covers repos where .symbiot isn't ignored.
+  const add = git(path, ["add", "-A"]); if (!add.ok) return { error: "git add failed: " + add.err, branch };
+  git(path, ["reset", "-q", "--", ".symbiot"]); // never ship Symbiot's own scratch
   const subject = texts.length === 1 ? (texts[0].length > 72 ? texts[0].slice(0, 71).replace(/\s+\S*$/, "") + "…" : texts[0]) : `symbiot: ${texts.length} approved tasks`;
   const body = texts.map((x) => "- " + x).join("\n");
   const cm = git(path, ["commit", "-m", subject, "-m", body]); if (!cm.ok) return { error: "Commit failed: " + (cm.err || cm.out), branch };
@@ -256,11 +261,7 @@ function pushTasks(filter) {
   if (filter && filter.type) tasks = tasks.filter((t) => taskType(t.text) === filter.type);
   if (filter && filter.repo) tasks = tasks.filter((t) => t.repo === filter.repo);
   if (!tasks.length) return { empty: true, written: [], unresolved: [] };
-  // Resolve repo name -> path from the already-scanned map when we have it
-  // (avoids a fresh full scan); fall back to a scan only if needed.
-  const byName = {};
-  const src = (LAST_MAP && LAST_MAP.nodes) ? LAST_MAP.nodes.filter((n) => n.type === "repo" || n.type === "folder").map((n) => ({ name: n.label, path: n.meta && n.meta.path })) : findAllRepos();
-  for (const r of src) if (r.path && !byName[r.name]) byName[r.name] = r.path;
+  const byName = repoPathMap(); // from the already-scanned map when there is one
   const groups = {}; for (const t of tasks) { const k = t.repo || ""; (groups[k] = groups[k] || []).push(t); }
   const written = [], unresolved = [];
   for (const name of Object.keys(groups)) {
@@ -717,44 +718,47 @@ function recommendModels(hw) {
   return { local, paid, best };
 }
 // ---- hand a repo (+ its tasks) to the user's agent — generic, settings-based
-// The handoff is a command TEMPLATE the user configures, with {dir} (repo path)
-// and {prompt} (the task instruction). Works for any agent/editor:
+// ONE code path, whatever the agent. The handoff is a command TEMPLATE the user
+// saves (`agentCmd`), with {dir} (repo path) and {prompt} (the task
+// instruction), and it is the only thing that ever runs:
+//   handoffCmd() -> fillHandoff() -> runHandoff() -> track()
+// Everything else in this section only helps pick a template: one-click presets
+// for what's installed (detectHandoffs), the Orca preset (a template too, kept
+// current by migrateOrcaCmd) and the cross-platform lookups behind them.
 //   claude -p "{prompt}"      · aider --message "{prompt}"      · code {dir}
 //   gnome-terminal --working-directory={dir} -- claude "{prompt}"
 // Not tied to any one tool — you decide what runs.
-// [cmd, label, macAppName] — macApp used to launch GUI editors on macOS where
-// the CLI isn't on PATH (they're .app bundles).
+const shSingle = (s) => "'" + String(s).replace(/'/g, "'\\''") + "'";
+const escDq = (s) => String(s).replace(/[\\"$`]/g, "\\$&");
+// Plain words only: in the Orca preset this passes through two shells, so no
+// backticks or $ (escDq covers one level).
+const HANDOFF_PROMPT = "Read .symbiot/TASKS.md and implement the unchecked items in this repo. Tick each item [x] in that file as you finish it and leave your changes uncommitted, so they can be reviewed and approved. Confirm before anything destructive.";
+function handoffCmd() { const cfg = loadConfig(); return migrateOrcaCmd(cfg.agentCmd) || (cfg.ide ? `${cfg.ide} {dir}` : ""); } // ide = legacy
+// Save the template ("" clears it). Either way the legacy `ide` key goes.
+function setHandoffCmd(cmd) {
+  const cfg = loadConfig(); const v = String(cmd || "").trim();
+  if (v) cfg.agentCmd = v; else delete cfg.agentCmd;
+  delete cfg.ide; saveConfig(cfg);
+  return { ok: true, cmd: cfg.agentCmd || "" };
+}
+const fillHandoff = (tmpl, repoPath) => tmpl.replace(/\{dir\}/g, shSingle(repoPath)).replace(/\{prompt\}/g, escDq(HANDOFF_PROMPT));
+function runHandoff(repoPath) {
+  const tmpl = handoffCmd(); if (!tmpl || !repoPath) return null;
+  return track(repoPath.split("/").pop(), fillHandoff(tmpl, repoPath), repoPath);
+}
+// Presets. [cmd, label, macAppName] — macApp used to launch GUI editors on macOS
+// where the CLI isn't on PATH (they're .app bundles).
 const IDE_LIST = [["code", "VS Code", "Visual Studio Code"], ["cursor", "Cursor", "Cursor"], ["windsurf", "Windsurf", "Windsurf"], ["zed", "Zed", "Zed"], ["subl", "Sublime Text", "Sublime Text"], ["idea", "IntelliJ IDEA", "IntelliJ IDEA"], ["nvim", "Neovim", ""]];
+// [cmd, label, template]. Only agents that leave changes to review: a handoff
+// runs unattended, so a "plan only" run can't ask anything and leaves nothing to
+// approve — that preset was dropped (a saved one still runs as-is).
 const AGENT_LIST = [
   ["claude", "Claude Code — make changes", 'claude -p "{prompt}" --permission-mode acceptEdits'],
-  ["claude", "Claude Code — plan only (asks first)", 'claude -p "{prompt}"'],
   ["codex", "Codex (OpenAI/GPT) — make changes", 'codex exec --full-auto "{prompt}"'],
   ["aider", "Aider — make changes", 'aider --message "{prompt}" --yes'],
   ["gemini", "Gemini — make changes", 'gemini --yolo -p "{prompt}"'],
   ["cursor-agent", "Cursor agent", 'cursor-agent -p "{prompt}"'],
 ];
-// Cross-platform "is this command available?" (command -v on posix, where on win).
-function hasCmd(cmd) { try { return !!sh(process.platform === "win32" ? `where ${cmd}` : `command -v ${cmd}`).trim(); } catch { return false; } }
-function macApp(name) { if (process.platform !== "darwin" || !name) return ""; for (const base of ["/Applications", join(homedir(), "Applications")]) { try { if (existsSync(join(base, name + ".app"))) return name; } catch {} } return ""; }
-// Find the Orca IDE CLI across OSes (known locations, then a bounded search).
-let ORCA_CLI; // cached per process: undefined=unchecked, ""=none, string=path
-function findOrcaCli() {
-  if (ORCA_CLI !== undefined) return ORCA_CLI;
-  ORCA_CLI = _findOrcaCli();
-  return ORCA_CLI;
-}
-function _findOrcaCli() {
-  const home = homedir(); const cands = [];
-  if (process.platform === "linux") cands.push(join(home, ".local/share/orca-ide/app/resources/bin/orca-ide"));
-  if (process.platform === "darwin") { cands.push(join(home, "Library/Application Support/orca-ide/app/resources/bin/orca-ide"), "/Applications/Orca.app/Contents/Resources/app/resources/bin/orca-ide", join(home, "Applications/Orca.app/Contents/Resources/app/resources/bin/orca-ide")); }
-  if (process.platform === "win32") { const la = process.env.LOCALAPPDATA || ""; cands.push(join(la, "orca-ide", "app", "resources", "bin", "orca-ide"), join(la, "Programs", "orca-ide", "resources", "app", "resources", "bin", "orca-ide")); }
-  for (const c of cands) { try { if (existsSync(c)) return c; } catch {} }
-  const roots = process.platform === "darwin" ? [join(home, "Library/Application Support"), "/Applications", join(home, "Applications")]
-    : process.platform === "win32" ? [process.env.LOCALAPPDATA || "", process.env.PROGRAMFILES || ""]
-    : [join(home, ".local/share"), "/opt", join(home, ".config")];
-  for (const r of roots) { if (!r) continue; const hit = sh(`find ${JSON.stringify(r)} -maxdepth 6 -name orca-ide -type f 2>/dev/null | head -1`).trim(); if (hit) return hit; }
-  return "";
-}
 function detectHandoffs() {
   const editors = [];
   for (const [cmd, label, app] of IDE_LIST) {
@@ -779,14 +783,8 @@ function detectHandoffs() {
   }
   return { agents, editors };
 }
-const shSingle = (s) => "'" + String(s).replace(/'/g, "'\\''") + "'";
-const escDq = (s) => String(s).replace(/[\\"$`]/g, "\\$&");
-// Plain words only: in the Orca preset this passes through two shells, so no
-// backticks or $ (escDq covers one level).
-const HANDOFF_PROMPT = "Read .symbiot/TASKS.md and implement the unchecked items in this repo. Tick each item [x] in that file as you finish it and leave your changes uncommitted, so they can be reviewed and approved. Confirm before anything destructive.";
-function handoffCmd() { const cfg = loadConfig(); return migrateOrcaCmd(cfg.agentCmd) || (cfg.ide ? `${cfg.ide} {dir}` : ""); } // ide = legacy
-// Build the full Orca handoff, cold-start safe. `bin` is the quoted orca-ide
-// path; `commandPart` is e.g. ORCA_CLAUDE_CMD or "" (open only).
+// The Orca preset. Build the full Orca handoff, cold-start safe. `bin` is the
+// quoted orca-ide path; `commandPart` is e.g. ORCA_CLAUDE_CMD or "" (open only).
 // open launches Orca & waits for the runtime to be REACHABLE, but on a cold
 // start the workspace graph isn't ready yet (runtime.state=graph_not_ready) and
 // `terminal create` times out — so we poll `status` until state=ready, then add
@@ -817,6 +815,30 @@ function migrateOrcaCmd(cmd) {
   if (rebuilt !== cmd) { try { const cfg = loadConfig(); if (cfg.agentCmd === cmd) { cfg.agentCmd = rebuilt; saveConfig(cfg); } } catch {} }
   return rebuilt;
 }
+// Cross-platform detection for the presets (and gh / the app's browser).
+// "Is this command available?": command -v on posix, where on win.
+function hasCmd(cmd) { try { return !!sh(process.platform === "win32" ? `where ${cmd}` : `command -v ${cmd}`).trim(); } catch { return false; } }
+function macApp(name) { if (process.platform !== "darwin" || !name) return ""; for (const base of ["/Applications", join(homedir(), "Applications")]) { try { if (existsSync(join(base, name + ".app"))) return name; } catch {} } return ""; }
+// Find the Orca IDE CLI across OSes (known locations, then a bounded search).
+let ORCA_CLI; // cached per process: undefined=unchecked, ""=none, string=path
+function findOrcaCli() {
+  if (ORCA_CLI !== undefined) return ORCA_CLI;
+  ORCA_CLI = _findOrcaCli();
+  return ORCA_CLI;
+}
+function _findOrcaCli() {
+  const home = homedir(); const cands = [];
+  if (process.platform === "linux") cands.push(join(home, ".local/share/orca-ide/app/resources/bin/orca-ide"));
+  if (process.platform === "darwin") { cands.push(join(home, "Library/Application Support/orca-ide/app/resources/bin/orca-ide"), "/Applications/Orca.app/Contents/Resources/app/resources/bin/orca-ide", join(home, "Applications/Orca.app/Contents/Resources/app/resources/bin/orca-ide")); }
+  if (process.platform === "win32") { const la = process.env.LOCALAPPDATA || ""; cands.push(join(la, "orca-ide", "app", "resources", "bin", "orca-ide"), join(la, "Programs", "orca-ide", "resources", "app", "resources", "bin", "orca-ide")); }
+  for (const c of cands) { try { if (existsSync(c)) return c; } catch {} }
+  const roots = process.platform === "darwin" ? [join(home, "Library/Application Support"), "/Applications", join(home, "Applications")]
+    : process.platform === "win32" ? [process.env.LOCALAPPDATA || "", process.env.PROGRAMFILES || ""]
+    : [join(home, ".local/share"), "/opt", join(home, ".config")];
+  for (const r of roots) { if (!r) continue; const hit = sh(`find ${JSON.stringify(r)} -maxdepth 6 -name orca-ide -type f 2>/dev/null | head -1`).trim(); if (hit) return hit; }
+  return "";
+}
+// ---- background jobs --------------------------------------------------------
 // Run a shell command as a tracked, logged background job that shows up live in
 // the Agents tab. Shared by the agent handoff and the local-model setup.
 function track(name, cmd, cwd, onExit) {
@@ -851,11 +873,6 @@ function agentChanges(path, startedAt) {
     return { dirty, stat, commits };
   } catch { return { dirty: 0, stat: "", commits: [] }; }
 }
-const fillHandoff = (tmpl, repoPath) => tmpl.replace(/\{dir\}/g, shSingle(repoPath)).replace(/\{prompt\}/g, escDq(HANDOFF_PROMPT));
-function runHandoff(repoPath) {
-  const tmpl = handoffCmd(); if (!tmpl || !repoPath) return null;
-  return track(repoPath.split("/").pop(), fillHandoff(tmpl, repoPath), repoPath);
-}
 // ---- local model one-command setup (Ollama), OS-aware ---------------------
 const sleep = (ms) => new Promise((r) => setTimeout(r, ms));
 function hasOllama() { return !!sh("ollama --version").trim(); } // cross-platform (not POSIX command -v)
@@ -923,7 +940,7 @@ function worktreeCount(repo) { return sh(`git -C ${JSON.stringify(repo)} worktre
 // repo file, which could be attacker-controlled): ~/.config/symbiot/deploys.json
 function loadDeploys() { try { return JSON.parse(readFileSync(join(CONFIG_DIR, "deploys.json"), "utf8")); } catch { return {}; } }
 function ciState(repo, def) {
-  if (!sh("command -v gh 2>/dev/null").trim()) return null;
+  if (!hasCmd("gh")) return null;
   if (!/github\.com/i.test(sh(`git -C ${JSON.stringify(repo)} remote get-url origin 2>/dev/null`))) return null;
   const j = sh(`cd ${JSON.stringify(repo)} && gh run list --branch ${def} --limit 1 --json databaseId,conclusion,status 2>/dev/null`).trim();
   if (!j) return null;
@@ -1472,7 +1489,6 @@ function readBody(req) {
     req.on("end", () => { try { resolve(d ? JSON.parse(d) : {}); } catch { resolve({}); } });
   });
 }
-function onPath(cands) { for (const cc of cands) { if (sh(`command -v ${cc} 2>/dev/null`).trim()) return cc; } return null; }
 // Find a Chromium-family browser for the chrome-less --app window, per OS.
 function chromeBinary() {
   const p = process.platform;
@@ -1491,7 +1507,7 @@ function chromeBinary() {
     const w = sh("where chrome 2>NUL").split(/\r?\n/).map((s) => s.trim()).find(Boolean);
     return w || null;
   }
-  return onPath(["google-chrome", "google-chrome-stable", "chromium", "chromium-browser", "brave-browser", "microsoft-edge"]);
+  return ["google-chrome", "google-chrome-stable", "chromium", "chromium-browser", "brave-browser", "microsoft-edge"].find(hasCmd) || null;
 }
 function openApp(url) {
   try {
@@ -1544,7 +1560,7 @@ async function cmdApp() {
       if (u.pathname === "/api/scanroots/add" && req.method === "POST") { const b = await readBody(req); return json(res, addScanRoot(String(b.path || ""))); }
       if (u.pathname === "/api/scanroots/remove" && req.method === "POST") { const b = await readBody(req); return json(res, removeScanRoot(String(b.path || ""))); }
       if (u.pathname === "/api/agentcfg") { const d = detectHandoffs(); return json(res, { cmd: handoffCmd(), agents: d.agents, editors: d.editors }); }
-      if (u.pathname === "/api/agentcmd" && req.method === "POST") { const b = await readBody(req); const cfg = loadConfig(); const v = String(b.cmd || "").trim(); if (v) cfg.agentCmd = v; else delete cfg.agentCmd; delete cfg.ide; saveConfig(cfg); return json(res, { ok: true, cmd: cfg.agentCmd || "" }); }
+      if (u.pathname === "/api/agentcmd" && req.method === "POST") { const b = await readBody(req); return json(res, setHandoffCmd(b.cmd)); }
       if (u.pathname === "/api/open" && req.method === "POST") { const b = await readBody(req); const e = runHandoff(String(b.path || "")); return json(res, { opened: !!e, id: e ? e.id : "" }); }
       if (u.pathname === "/api/setup-local" && req.method === "POST") {
         const b = await readBody(req);
@@ -1582,7 +1598,7 @@ async function cmdApp() {
   server.on("listening", () => {
     if (announced) return; announced = true;
     const url = `http://127.0.0.1:${server.address().port}/?t=${TOKEN}`;
-    const how = opened ? "" : openApp(url); opened = true; // only pop a window the first time
+    const how = opened || process.env.SYMBIOT_NO_OPEN === "1" ? "" : openApp(url); opened = true; // only pop a window the first time (never in tests)
     console.log(`\n${c.g("●")} ${c.b("Symbiot")} is running at ${c.b(url)}`);
     console.log(how ? c.d(`  Opened in a ${how}.`) : c.d("  Open that URL in your browser."));
     console.log(c.d("  Leave this running; press Ctrl+C to stop (or click Quit in the window)."));
@@ -1605,7 +1621,7 @@ ${c.b("Usage")}
   symbiot todo                      what's still on your plate
   symbiot app                       open the visual app in your browser
   symbiot drift                     what's out of sync / at risk across repos
-  symbiot push [--open]             write tasks into each repo (and open your IDE)
+  symbiot push [--open]             write tasks into each repo (and run your agent)
   symbiot models                    recommend AI models for your hardware
   symbiot setup-local [--model X]   install/run a free local model (Ollama)
   symbiot login                     connect it to an AI (once)
@@ -1655,4 +1671,4 @@ const isMain = (() => {
 })();
 if (isMain) main();
 
-export { authorship, repoState, readmeInfo, repoShape, houseRules, findAllRepos, buildMap, reportFooter, detectHardware, recommendModels, computeDrift, driftRepo, gitDefaultBranch, buildTasksMd, taskType, EMBEDDED_UI, orcaHandoffCmd, migrateOrcaCmd, fillHandoff, ORCA_CLAUDE_CMD, HANDOFF_PROMPT, shipChanges, syncTasks, pendingReview, approveRepo, sendBack, pushTasks, semverGt };
+export { authorship, repoState, readmeInfo, repoShape, houseRules, findAllRepos, buildMap, reportFooter, detectHardware, recommendModels, computeDrift, driftRepo, gitDefaultBranch, buildTasksMd, taskType, EMBEDDED_UI, orcaHandoffCmd, migrateOrcaCmd, fillHandoff, handoffCmd, setHandoffCmd, ORCA_CLAUDE_CMD, HANDOFF_PROMPT, shipChanges, syncTasks, pendingReview, approveRepo, sendBack, pushTasks, semverGt };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "symbiot",
-  "version": "0.28.2",
+  "version": "0.28.4",
   "description": "Your week, written from your real work. Reads your local git and writes it up with the AI of your choice — Claude, OpenAI, Gemini, or a local model.",
   "type": "module",
   "bin": {
@@ -43,6 +43,6 @@
     "@anthropic-ai/sdk": "^0.128.0"
   },
   "scripts": {
-    "test": "node test/fixtures.mjs && node test/smoke.mjs && node test/install.mjs"
+    "test": "node test/load.mjs && node test/fixtures.mjs && node test/smoke.mjs && node test/install.mjs"
   }
 }

package/ui.mjs

@@ -149,7 +149,7 @@ footer{padding:10px 18px;border-top:1px solid var(--line);display:flex}
 <div class="rfoot" id="outfoot" style="display:none"></div>
 </section>
 <section id="panel-tasks" class="hidden">
-<div class="row"><input id="newtask" placeholder="Add a task..." style="flex:1"><button class="act" id="addtask">Add</button><button class="ghost" id="pushtasks" title="Write .symbiot/TASKS.md into each repo for your coding agent">Send to repos</button></div>
+<div class="row"><input id="newtask" placeholder="Add a task..." style="flex:1"><select id="newtaskrepo" title="Which repo this task is for (needed to send it to an agent)" style="flex:0 0 auto;max-width:180px"><option value="">repo…</option></select><button class="act" id="addtask">Add</button><button class="ghost" id="pushtasks" title="Write .symbiot/TASKS.md into each repo for your coding agent">Send to repos</button></div>
 <div class="note muted" style="margin-top:2px">To give tasks to your agent, use <b>Send to repos</b> &mdash; filter by tag below to choose which. As the agent finishes each one it lands in <b>Awaiting your review</b>: <b>Approve</b> commits it on a branch and opens a PR, <b>&#8630;</b> sends it back. Ticking a task yourself marks it done (it auto-archives).</div>
 <div id="pushout"></div>
 <div id="reviewout"></div>
@@ -192,7 +192,7 @@ footer{padding:10px 18px;border-top:1px solid var(--line);display:flex}
 </div>
 <div style="margin-top:20px;border-top:1px solid var(--line);padding-top:16px">
 <label>Hand off to your agent when you "Send to repos"</label>
-<input id="agentcmd" type="text" placeholder="e.g.  claude -p &quot;{prompt}&quot;   ·   code {dir}   ·   leave blank to just write the file">
+<input id="agentcmd" type="text" placeholder="e.g.  aider --message &quot;{prompt}&quot; --yes   ·   code {dir}   ·   leave blank to just write the file">
 <div class="note muted">Runs in each repo after tasks are written. Use <b>{dir}</b> = repo path, <b>{prompt}</b> = the task instruction. Works with any agent or editor &mdash; it's your command.</div>
 <div id="agentpresets" style="margin-top:8px"></div>
 </div>
@@ -224,7 +224,7 @@ $('panel-drift').classList.toggle('hidden',!isDrift);
 $('panel-agents').classList.toggle('hidden',!isAgents);
 if(isRun){$('what').textContent=tab;$('out').textContent='Nothing yet - hit the button.';$('out').classList.add('muted');$('copy').classList.add('hidden');}
 if(isMap&&!mapLoaded)loadMap();
-if(isTasks)loadTasks();
+if(isTasks){fillTaskRepos();loadTasks();}
 if(isDrift&&!driftLoaded)loadDrift();
 if(isAgents)loadAgents(); else stopAgentsPoll();}
 tabs().forEach(function(t){t.addEventListener('click',function(){setTab(t.dataset.tab);});});
@@ -377,14 +377,15 @@ var clean=repos.filter(function(r){return !r.flags.length;}).map(function(r){ret
 if(clean.length)h+="<div class='muted' style='margin-top:10px'>clean: "+esc(clean.join(", "))+"</div>";
 if(!repos.length)h+="<div class='muted'>No repos found under your home folder.</div>";
 out.innerHTML=h;});}
-function addTaskUI(){var i=document.getElementById('newtask');var v=(i.value||'').trim();if(!v)return;api('/api/tasks/add',{text:v,repo:''}).then(function(){i.value='';loadTasks();});}
+function fillTaskRepos(){var sel=document.getElementById('newtaskrepo');if(!sel||!GRAPH)return;var cur=sel.value;var names=GRAPH.nodes.filter(function(n){return n.type==='repo'||n.type==='folder';}).map(function(n){return n.label;}).sort();sel.innerHTML="<option value=''>repo…</option>"+names.map(function(n){return "<option value='"+esc(n)+"'"+(n===cur?" selected":"")+">"+esc(n)+"</option>";}).join("");}
+function addTaskUI(){var i=document.getElementById('newtask');var v=(i.value||'').trim();if(!v)return;var repo=(document.getElementById('newtaskrepo')||{}).value||'';api('/api/tasks/add',{text:v,repo:repo}).then(function(){i.value='';TFILTER={type:'',repo:''};loadTasks();});}
 function pushTasksUI(){var o=document.getElementById('pushout');var btn=document.getElementById('pushtasks');btn.disabled=true;
 o.innerHTML="<div class='muted' style='margin-top:10px'><span class='dot-c' style='background:var(--amber)'></span> Writing .symbiot/TASKS.md into your repos&hellip;</div>";
 api('/api/tasks/push',{type:TFILTER.type||'',repo:TFILTER.repo||''}).then(function(r){btn.disabled=false;
 if(r.empty){o.innerHTML="<div class='muted' style='margin-top:10px'>No open tasks to send. Tick ideas in a repo review, or add tasks above.</div>";return;}
 var n=(r.written||[]).length;var h="<div class='drift' style='margin-top:10px'><div class='dh'><span class='dot-c'></span><span class='dn'>Done &mdash; wrote "+n+" file"+(n===1?"":"s")+"</span></div>";
 if(n){h+="<ul>";r.written.forEach(function(w){h+="<li class='info'>&#10003; <b>"+esc(w.name)+"</b> <span class='ev'>"+esc(w.file)+" ("+w.count+" task"+(w.count===1?"":"s")+")</span></li>";});h+="</ul>";}
-if(r.unresolved&&r.unresolved.length){h+="<div class='dd' style='margin-top:6px'>&#9888; not written (repo not found under your home folder): "+esc(r.unresolved.map(function(u){return u.name;}).join(", "))+"</div>";}
+if(r.unresolved&&r.unresolved.length){var names=r.unresolved.map(function(u){return u.name;}).join(", ");var hasNoRepo=r.unresolved.some(function(u){return u.name==='(no repo)';});h+="<div class='dd' style='margin-top:6px'>&#9888; not sent: "+esc(names)+". "+(hasNoRepo?"Pick a repo in the dropdown next to <b>Add</b> so the task has somewhere to go.":"That repo isn't in the map &mdash; add its folder in Settings.")+"</div>";}
 if(r.handoff&&r.written&&r.written.length){r.written.forEach(function(w){api('/api/open',{path:w.path});});h+="<div class='dd' style='margin-top:8px'>&#129302; Handed "+n+" repo(s) to your agent &mdash; opening the <b>Agents</b> tab to watch it work&hellip;</div>";setTimeout(function(){setTab('agents');},500);}
 else{h+="<div class='dd' style='margin-top:8px'>Set an <b>agent command</b> in Settings to auto-run it on send (and watch it in the Agents tab). For now, tell your agent: <b>“Read .symbiot/TASKS.md and implement the unchecked items.”</b></div>";}
 h+="</div>";
@@ -423,7 +424,7 @@ function scanLine(s){return 'Mapping your work… '+s.phase+(s.total?' '+s.done+
 function loadMap(){var p=document.getElementById("profile");p.textContent="Mapping your work...";document.getElementById("graph").innerHTML="";sel=null;hideDetail();hideReview();var done=false;
 function poll(){if(done)return;api('/api/scan').then(function(s){if(done)return;if(s&&s.active&&s.phase)p.textContent=scanLine(s);setTimeout(poll,600);}).catch(function(){});}
 setTimeout(poll,400);
-api("/api/map").then(function(g){done=true;mapLoaded=true;if(!g.nodes||!g.nodes.length){p.textContent="No git repositories found under your home folder.";return;}GRAPH=g;layout(g.nodes,g.edges);view={k:1,x:0,y:0};p.innerHTML=profileLine(g)+(g.stats&&g.stats.partial?" &middot; <span class='err'>partial &mdash; the scan hit its time limit</span>":"");render();});}
+api("/api/map").then(function(g){done=true;mapLoaded=true;if(!g.nodes||!g.nodes.length){p.textContent="No git repositories found under your home folder.";return;}GRAPH=g;fillTaskRepos();layout(g.nodes,g.edges);view={k:1,x:0,y:0};p.innerHTML=profileLine(g)+(g.stats&&g.stats.partial?" &middot; <span class='err'>partial &mdash; the scan hit its time limit</span>":"");render();});}
 function fitBadge(m){return m.fits?"<span class='tag'>fits your RAM</span>":"<span class='tag' style='background:#3a2a12;border-color:#6b4a1f;color:#F2A541'>needs more RAM</span>";}
 function loadRec(){var out=document.getElementById('recout');out.innerHTML="<div class='muted'>Reading your hardware...</div>";
 api('/api/models').then(function(d){var hw=d.hardware,rec=d.rec;