@namzu/cli 36.2.0 → 36.3.1

package/CHANGELOG.md

@@ -1,5 +1,46 @@
 # @namzu/cli
 
+## 36.3.1
+
+### Patch Changes
+
+- 6f80b14: The `workflow` tool now reports a script that fails before it starts any agent as a failed tool call with the located error (line, column, source line), instead of "started in the background" followed by a separate failure message and a dead `0/0` row on the rail; a script that finishes without an agent returns its result in the call. `wait_for_task` given a workflow run id now says in a sentence that a workflow run is not a task and that its end arrives as one message, instead of a UUID format error. Nothing for callers to change.
+
+## 36.3.0
+
+### Minor Changes
+
+- 437dd6d: Add the `workflow` tool. When you ask for one (hypermode on, the word "workflow" or "iş akışı" in your message, or `/workflow <what to do>`), the model can write a short script that starts many agents in phases and runs them in the background. Before anything starts you see its name, purpose, phases, agent limits and the script (`d` shows all of it) and approve once; in `auto` mode ("bypass permissions") it starts without asking, and a session with nobody to ask refuses with a plain reason. Its agents are ordinary children under the same permission mode, rules, sandbox and budget as the `Agent` tool, but they are not subagent rows: the run is one row on the rail (`◯ name ▰▱ 3/6 · 12s · ↓ tokens`) opening to its phases (declared ones listed before they start), then a phase's agents, then an agent's transcript; `x` stops the run on the phases and skips one agent on the agents. The conversation shows `● Workflow(<purpose>)` at launch and `✓ Workflow <name> completed · …` when the run ends, `/workflows` lists this conversation's runs (earlier ones from their journals under `<session>/runs/<runId>/workflow/`), and the model gets one message with the result when a run ends. Nothing changes for conversations that never ask for a workflow. The workflow tool is offered only by hosts that can review a background run (the terminal and `namzu exec`, not the desktop app yet). A finished workflow and a background command that ends on its own now start a turn at once when the session is idle, so the model reports without you typing (a command you stopped yourself does not). `namzu exec` waits for a workflow the model started, lets the model report, and then exits; its `Tokens:` line (and the `budget` in `exec --json`) is the whole command's, the reporting turn and the workflow's agents included. New config keys: `workflow.maxConcurrency` (1-16, default min(8, CPUs-2)) and `workflow.maxAgents` (1-1000, default 200).
+
+### Patch Changes
+
+- 437dd6d: The agents rail under the message box is now a `main` row plus one row per agent that is running or finished in the last 30 seconds (type, description, time and spend), five at a time with `↓ N more`. Press ↓ in an empty composer to move into it: ↑/↓ select, Enter opens that agent's transcript (Esc comes back to the same row), `x` stops a running agent or clears a finished one, Esc returns to the composer. The old titled tree (workflow and phase lines, a latest-activity line per agent, the `↓ / ctrl+t` counter) is no longer drawn on the rail; `Ctrl+T` and `/agents` still open the full cockpit with all of it.
+- 437dd6d: `/jobs` is now the `Background` panel: a row per background shell (status icon, command, `running` / `done` / `error` / `stopped`), `Shell details` with `Status:`, `Runtime:`, `Command:` and live `Output:`, and no job ids on screen. When a background command ends the transcript gets a row such as `● Background command "npm test" completed (exit code 0) · 12s`, or `failed with exit code 2` / `was stopped`.
+- 437dd6d: Messages you type while Namzu is working now wait as grey `❯` rows above the input box instead of inside it; a Tab-queued one says `after this reply`. Press Up on an empty input to choose a queued message and Enter to edit it, or `ctrl+x ctrl+s` to send waiting messages now. Nothing changes in how messages reach the model.
+
+  They sit directly above the `Working` row.
+
+- 692466e: The message box in Namzu Desktop now shows, beside the mode, how many background shells are running ("2 shells", which opens the list) and the pull request of the conversation's branch ("PR #607", coloured by review state, which opens it in the browser). The pull request comes from the GitHub command line tool the terminal already uses; with `gh` missing, signed out, or no pull request, nothing is shown. The CLI host gains a read-only `namzu/project/pr` request for it.
+- 692466e: A message typed in Namzu Desktop while a reply is running now reaches that reply at its next model request, as the terminal does. Before, the desktop host reported live input as unavailable for the whole of every turn, so each such message silently waited for the reply to finish. Nothing to change on your side.
+- 437dd6d: The composer footer is now one row of items in the reference's order: the mode in plain words (`⏵⏵ bypass permissions on`, `⏵⏵ accept edits on`, `‖ plan mode on`), `PR #592` for the current branch's pull request (coloured by review state and linked where the terminal supports it), a `N shells` pill for running background shells (the separate "shells running" line is gone), then `esc to interrupt` while a turn runs, `← N agents` and `↓ to manage`. From an empty composer `←` opens the agents and `↓` focuses the shell pill first (Enter opens the `/jobs` screen) and then the agents rail; with anything typed both keys keep moving the cursor, and the hints leave the row. The agents rail's own `↓ to manage` on its `main` row is gone, since the footer says it. A running shell takes the `(shift+tab to cycle)` reminder off the row. The model name now yields first when the row is narrow. The pull request comes from the `gh` command line tool, read in the background, only on a real terminal; `NAMZU_PR_STATUS=off` turns it off. The working directory no longer appears on the footer row. The composer's frame is now two plain horizontal rules with a dim `❯ ` prompt between them: the boxed frame with corners, side borders and a `MESSAGE` caption is gone, and `hypermode` is named at the upper rule's right end.
+- 437dd6d: The startup header now shows the folder Namzu is working in on the row under the version, with the home folder written as `~` and a long path cut from the left to fit. The footer no longer carries it.
+- 437dd6d: Shell commands in the terminal now read like Claude Code's: a green or red dot in front of `Bash(command)` instead of a tick and a duration, the command's line breaks kept in the header, and the output hung under `⎿` as its first three lines followed by `… +N lines (ctrl+o to expand)`. A failed command now shows its error output and `Exit code N` instead of repeating the command, an empty result reads `(No output)`, and a command moved to a background job reads `Running in the background` instead of the instructions meant for the model. While a command runs it shows `Running…` with the elapsed time.
+- 28eaee1: On Windows the `bash` tool, background jobs started with `run_in_background`, and plugin shell hooks now run in Git for Windows' bash instead of cmd.exe whenever Git for Windows is installed, and `hostCommandShell()` and `BashTool.commandDialect()` answer `bash` there instead of `cmd`. Before, every command a model wrote was refused or needed a person to approve it each time, because the permission rules cannot read cmd.exe lines. Git Bash is looked up as `NAMZU_BASH_SHELL`, then `%ProgramFiles%\Git\bin\bash.exe`, then `%ProgramFiles(x86)%\Git\bin\bash.exe`, then `%LOCALAPPDATA%\Programs\Git\bin\bash.exe`, then the install holding the first `git` that `where.exe git` lists outside the working directory. Commands now need Unix syntax (`/dev/null`, forward slashes, `$VAR`); the model is told so, and `>nul` is rewritten to `/dev/null`. A host that depends on cmd.exe syntax in the `bash` tool, or on `hostCommandShell().dialect === 'cmd'` on a machine with Git for Windows, sets `NAMZU_BASH_SHELL=cmd` (a path to `cmd.exe`, such as `%ComSpec%`, works too) to keep the old behaviour. Without Git for Windows nothing changes except that `windowsShellNotice()` returns a sentence saying commands run in cmd.exe and each asks for approval; the CLI shows it once at session start. New exports: `findGitBash`, `isWindowsBash`, `windowsShellNotice`, `rewriteWindowsNullRedirect`, `WINDOWS_BASH_GUIDANCE`, `WINDOWS_CMD_FALLBACK_NOTICE`, and an optional `windows` flag on `CommandShell`. Stopping a command or job on Windows now also ends processes that Git Bash forked, which `taskkill /T` alone missed.
+- 437dd6d: Agents that the model gave one `workflow` label now also appear as a single row on the rail under the message box: `◯ name`, a progress bar, `3/7 · elapsed · ↓ tokens`. Press ↓ to reach it; Enter opens that workflow's phases and agents (Esc comes back to the row), `x` stops its running agents or clears the row once it has finished, and the row leaves 30 seconds after the last agent ends. When a workflow's turn ends, one line in the conversation says how it ended. The rail's key line now reads `enter to view · x to stop`.
+- 437dd6d: Workflows and subagents are now separate in the terminal. An agent started with a `workflow` label no longer shows up as an agent row on the rail: it appears only inside its workflow, which is one row after the agent rows, and the `main` row shows only when there is an agent row. Enter on the workflow opens its phases (numbered, turning ✔ or ✘, with done/total), then the agents of a phase, then an agent's transcript; ← or Esc goes back up and `x` stops the whole workflow or one agent. In the conversation a workflow's launch is one `Workflow <name>` line instead of a row per agent, and the turn summary no longer counts phases.
+- Updated dependencies [437dd6d]
+- Updated dependencies [437dd6d]
+- Updated dependencies [28eaee1]
+- Updated dependencies [437dd6d]
+  - @namzu/sdk@51.0.0
+  - @namzu/browser@2.0.1
+  - @namzu/computer-use@3.1.0
+  - @namzu/anthropic@6.3.0
+  - @namzu/ollama@2.2.5
+  - @namzu/openai@4.1.4
+  - @namzu/openrouter@3.0.3
+  - @namzu/sandbox@26.0.0
+
 ## 36.2.0
 
 ### Minor Changes

package/dist/bin.js

@@ -81,6 +81,6 @@ __name(suppressSqliteExperimentalWarning, "suppressSqliteExperimentalWarning");
 suppressSqliteExperimentalWarning();
 
 // src/bin.ts
-var { main } = await import("./chunks/bin-main-FPJQMK33.js");
+var { main } = await import("./chunks/bin-main-J5ULRXXN.js");
 main(fileURLToPath(new URL("./", import.meta.url).href + "bin.js"));
 //# sourceMappingURL=bin.js.map

package/dist/chunks/{acp-P22WRX3A.js → acp-CJ2TTGJD.js}

@@ -2,15 +2,17 @@ import { createRequire as __namzuCreateRequire } from 'node:module';
 const require = __namzuCreateRequire(import.meta.url);
 import {
   dispatchCliPalMessages
-} from "./chunk-YLIKDPD6.js";
+} from "./chunk-S6DK2YOB.js";
 import {
   conversationMarkdown,
+  lookupPr,
+  runCommand,
   withKeyNote
-} from "./chunk-OS7IVIYH.js";
-import "./chunk-43FBWINL.js";
+} from "./chunk-CXSUA2PF.js";
+import "./chunk-AEPAIDAA.js";
 import {
   decideHeadlessTrust
-} from "./chunk-WNMRETWS.js";
+} from "./chunk-VUOS5SQ3.js";
 import {
   cliPalComputerStatus,
   cliPalScreen,
@@ -23,21 +25,21 @@ import {
   startCliPalComputer,
   stopCliPalComputer,
   takeOverCliPalComputer
-} from "./chunk-RPAHIOW6.js";
+} from "./chunk-YYQFUKI7.js";
 import {
   palSessionEnvironment
-} from "./chunk-7OEK2NAO.js";
+} from "./chunk-UXGI3FE5.js";
 import {
   cliPalActivitySubscriptionPolicy,
   cliPalActivitySubscriptionStore,
   cliPalCommunicationPolicy,
   cliPalCommunicationStore
-} from "./chunk-KNNQHRZ4.js";
+} from "./chunk-46B7KRXY.js";
 import {
   claimPalConversation,
   listPalConversations,
   palConversationBinding
-} from "./chunk-GWOW3U2S.js";
+} from "./chunk-6TDJXZ5W.js";
 import {
   FileCheckpointStore,
   PluginSettingsStore,
@@ -48,16 +50,16 @@ import {
   sessionLogCheckpointView,
   validateComposerSendSettings,
   verifyCredential
-} from "./chunk-H7VJHF5S.js";
-import "./chunk-PT7TZNBV.js";
-import "./chunk-LIZPXS73.js";
+} from "./chunk-H6MWKPHM.js";
+import "./chunk-52OVP5BG.js";
+import "./chunk-3F5DGROG.js";
 import "./chunk-WHODOMXA.js";
-import "./chunk-HIZ7F6BT.js";
-import "./chunk-YTJO7G24.js";
-import "./chunk-L6FMG3R6.js";
-import "./chunk-XAKL7C7B.js";
-import "./chunk-HJ42RG5U.js";
-import "./chunk-PJ3UYK2K.js";
+import "./chunk-JCDTR2ZU.js";
+import "./chunk-XCWZZZOW.js";
+import "./chunk-74KMYYM2.js";
+import "./chunk-AP5KD43Y.js";
+import "./chunk-5URT3ZM2.js";
+import "./chunk-MJO6CH5F.js";
 import "./chunk-ZWA2JRJV.js";
 import "./chunk-NUGPLXVK.js";
 import {
@@ -67,7 +69,7 @@ import {
   resolveTrustedProjectContext
 } from "./chunk-CHIEVVQ2.js";
 import "./chunk-GNBMKISG.js";
-import "./chunk-U5AWHJ7O.js";
+import "./chunk-MBBFX64S.js";
 import "./chunk-US5ZFPAU.js";
 import {
   automaticKind,
@@ -75,7 +77,7 @@ import {
   discoverProviders,
   hasApiCredential,
   isAnonymousZen
-} from "./chunk-CFN7CUD3.js";
+} from "./chunk-DYUOXP5Q.js";
 import {
   ALL_PROVIDER_IDS,
   PROVIDER_REGISTRY
@@ -83,13 +85,13 @@ import {
 import {
   compilePermissions,
   warnLegacyMcpPermissionNames
-} from "./chunk-DYBUCWAB.js";
-import "./chunk-YZ675PG4.js";
+} from "./chunk-QXDKRJZ5.js";
+import "./chunk-HSEJGSXE.js";
 import {
   cliLogger
-} from "./chunk-2VXA7GZU.js";
+} from "./chunk-FU4GQ7K2.js";
 import "./chunk-EEDZOGV4.js";
-import "./chunk-AGHVC2IJ.js";
+import "./chunk-LBH5WVXU.js";
 import "./chunk-OH6TB27S.js";
 import {
   isOfferableModel
@@ -112,11 +114,11 @@ import {
   refreshIndex,
   setTitle,
   unarchiveConversation
-} from "./chunk-ABD2DOO2.js";
+} from "./chunk-EKUJUPFA.js";
 import {
   cliProjectRoot
 } from "./chunk-H6B6AROX.js";
-import "./chunk-WW4VKEBD.js";
+import "./chunk-S6MZK4LD.js";
 import {
   cliPalStore,
   createPal,
@@ -126,13 +128,13 @@ import {
   listPals,
   palAtWorkspace,
   updatePal
-} from "./chunk-NRML7BIR.js";
+} from "./chunk-AS7UWXMN.js";
 import {
   isTrusted,
   isTrustedAtStateRoot,
   trustDir,
   untrustDir
-} from "./chunk-GFSD2D3M.js";
+} from "./chunk-ZG6NVROB.js";
 import {
   canonicalProjectPath
 } from "./chunk-JORONMD3.js";
@@ -143,8 +145,8 @@ import {
   readStoredGeminiApiKey,
   writeStoredApiKey,
   writeStoredGeminiApiKey
-} from "./chunk-QQLB573I.js";
-import "./chunk-3OZ2OBGU.js";
+} from "./chunk-BOV53LZO.js";
+import "./chunk-HKDPP2JF.js";
 import {
   ACPServer,
   DiskLogMedium,
@@ -177,7 +179,7 @@ import {
   resolveWithinAnyReal,
   resolveWithinReal,
   selectAssistantText
-} from "./chunk-ZKTLXAQK.js";
+} from "./chunk-V3ZLK3XQ.js";
 import {
   __commonJS,
   __name,
@@ -9587,6 +9589,7 @@ function withCliHarnesses(base, directory, deps = createDefaults()) {
   const selections = /* @__PURE__ */ new Map();
   const records = /* @__PURE__ */ new Map();
   const reserving = /* @__PURE__ */ new Set();
+  const nativeTurns = /* @__PURE__ */ new Set();
   const catalogues = /* @__PURE__ */ new Map();
   const clock2 = /* @__PURE__ */ __name(() => (deps.now ?? Date.now)(), "clock");
   let closed = false;
@@ -9754,7 +9757,14 @@ function withCliHarnesses(base, directory, deps = createDefaults()) {
       let record2 = records.get(request.sessionId);
       try {
         const choice = await selectionFor(request.sessionId, cwd);
-        if (!choice) return await base.gateway.prompt(request);
+        if (!choice) {
+          nativeTurns.add(request.sessionId);
+          try {
+            return await base.gateway.prompt(request);
+          } finally {
+            nativeTurns.delete(request.sessionId);
+          }
+        }
         if (request.attachments?.length)
           throw new Error(
             "This engine connection does not support attachments yet. Your draft is retained."
@@ -9864,7 +9874,7 @@ function withCliHarnesses(base, directory, deps = createDefaults()) {
     ...liveInputStatus && liveInput ? {
       liveInputStatus: /* @__PURE__ */ __name(async (id2, scopeId, scope) => {
         if (closed) throw new Error("The connection is closed.");
-        if (records.has(id2) || selections.has(id2) || reserving.has(id2)) {
+        if (records.has(id2) || selections.has(id2) || reserving.has(id2) && !nativeTurns.has(id2)) {
           if (scopeId !== void 0)
             throw new Error("This live prompt scope is unavailable in this engine.");
           return { available: false, inputs: [] };
@@ -9872,7 +9882,7 @@ function withCliHarnesses(base, directory, deps = createDefaults()) {
         return liveInputStatus(id2, scopeId, scope);
       }, "liveInputStatus"),
       liveInput: /* @__PURE__ */ __name(async (id2, input, scope) => {
-        if (closed || records.has(id2) || selections.has(id2) || reserving.has(id2))
+        if (closed || records.has(id2) || selections.has(id2) || reserving.has(id2) && !nativeTurns.has(id2))
           throw new Error("Live input is unavailable in this engine.");
         return liveInput(id2, input, scope);
       }, "liveInput")
@@ -10917,6 +10927,14 @@ function createProjectChanges(options = {}) {
 }
 __name(createProjectChanges, "createProjectChanges");
 
+// src/commands/desktop-host-pr.ts
+async function projectPr(cwd, run = runCommand) {
+  const found = await lookupPr(run, cwd);
+  if (found.state.kind !== "pr") return null;
+  return { number: found.state.number, url: found.state.url, review: found.state.review };
+}
+__name(projectPr, "projectPr");
+
 // src/commands/desktop-host.ts
 function text3(params, key, max = 400) {
   const value = params[key];
@@ -11979,6 +11997,12 @@ function createDesktopHostExtensions(runtime, directory, publishedSessionCwd, re
       if (!isTrusted(cwd)) return null;
       return projectGit(cwd);
     }, "namzu/project/git"),
+    "namzu/project/pr": /* @__PURE__ */ __name(async (params = {}) => {
+      if (!params || typeof params !== "object" || Array.isArray(params) || Object.keys(params).length > 0)
+        throw new Error("Invalid project pull request request.");
+      if (!isTrusted(cwd)) return { pr: null };
+      return { pr: await projectPr(cwd) };
+    }, "namzu/project/pr"),
     "namzu/project/changes": /* @__PURE__ */ __name(async (params = {}) => {
       if (!params || typeof params !== "object" || Array.isArray(params) || Object.keys(params).length > 0)
         throw new Error("Invalid project changes request.");
@@ -13516,4 +13540,4 @@ export {
   resolveAcpSession,
   runAcpCommand
 };
-//# sourceMappingURL=acp-P22WRX3A.js.map
+//# sourceMappingURL=acp-CJ2TTGJD.js.map