@cabane/companion 0.6.197 → 0.6.199

package/README.md

@@ -4,7 +4,7 @@ The Cabane Companion runs your Cabane agents on your own device. Install it, pas
 
 It drives a **harness** you already have signed in: [Claude Code](https://www.npmjs.com/package/@anthropic-ai/claude-code), [Codex](https://developers.openai.com/codex/cli) or [OpenCode](https://opencode.ai). The Companion never installs a harness and never runs a sign-in; it finds the ones that are usable and reports them.
 
-Step-by-step guides with screenshots are in Cabane: open the **Account menu** and choose **Docs**. This README is the reference for the device side.
+Step-by-step guides with screenshots are at [cabane.ai/docs](https://cabane.ai/docs): [Connect your computer](https://cabane.ai/docs/connect-your-computer) to install and pair, [Create an agent](https://cabane.ai/docs/create-an-agent) to give an agent a device, and [When an agent doesn't answer](https://cabane.ai/docs/when-an-agent-doesnt-answer) when one stays silent. This README is the reference for the device side.
 
 ## The model
 
@@ -131,7 +131,7 @@ Cabane has no sandbox setting of its own. A harness's own sandbox settings restr
 
 So where agents run is your choice, and there are two ways:
 
-- **The Companion in a container or VM**, recommended for almost everyone: the Companion and the harness together in Docker, OrbStack or a VM, sharing only the folders and credentials they need. Agents keep a full shell to work and code in, still with no approvals, while the container or VM limits what they reach to what you put in, MCP servers, hooks and the harness process included. If you haven't set one up before, ask Claude to help; **Common device patterns** in Cabane's Docs walks through one with Docker.
+- **The Companion in a container or VM**, recommended for almost everyone: the Companion and the harness together in Docker, OrbStack or a VM, sharing only the folders and credentials they need. Agents keep a full shell to work and code in, still with no approvals, while the container or VM limits what they reach to what you put in, MCP servers, hooks and the harness process included. If you haven't set one up before, ask Claude to help; [Common device patterns](https://cabane.ai/docs/common-device-patterns) walks through one with Docker.
 - **Directly on your device**, with the harness's own sandbox set to taste. It restricts particular tools, and each harness draws a different line.
 
 Running directly, confining agents is set in each harness's own file, and governs your own sessions there as well as every agent on the device:

package/dist/cli.js

@@ -2616,7 +2616,7 @@ var init_descriptions = __esm({
       "withdraw_ask"
     ];
     TURN_CONTROL_TOOL_DESCRIPTIONS = {
-      ask: `Ask a person a structured question you need answered to continue, then end your turn; don't wait for the reply. \`targetUserId\` is a person on your roster, by their \`@handle\`; to bring in an agent, name it with \`@handle\` in your reply.
+      ask: `Ask a person a structured question you need answered to continue, then end your turn; don't wait for the reply. \`targetUserId\` is a person on your roster, by their \`@handle\`; to bring in an agent, name them with \`@handle\` in your reply.
 - Form: one question (\`headline\`, optional \`question\` body, \`options\`) or \`questions\`, 1\u2013${MAX_ASK_ITEMS} \`{ headline, body?, options? }\`; never both. Decisions that arise together go in ONE call, in the order you want them answered. A second \`ask\` is a separate item, never an edit; raise one only for a genuinely new question.
 - \`headline\`: one sentence ending in \`?\`. Body (\`question\` or each item's \`body\`): a sentence or two of framing, not a report; inline Markdown only, no lists or headings. \`options\`: ${MIN_ASK_OPTIONS}\u2013${MAX_ASK_OPTIONS} one-click answers (always for a yes/no go-ahead), each a distinct course of action you can carry out without asking back.
 - Voice, in headline, body and options: "I" for yourself, never your own name; name everyone else who decides or acts, as on your roster. Never "you", "your", "we", "us", "our" or their contractions, nor "he", "she" or "they" for a named actor. Each option reads correctly alone. A quotation stays accurate: say who said it.
@@ -12686,7 +12686,7 @@ function turnFailureCopy(reason, runtime) {
   if (reason === "timeout_total")
     return "The turn ran longer than expected and was stopped. Reply in Cabane to continue.";
   if (reason === "result_error:error_max_turns")
-    return "The agent reached its step limit. Reply in Cabane to continue.";
+    return "The agent reached their step limit. Reply in Cabane to continue.";
   if (reason === "no_result" || reason === "empty_result" || reason === "empty_result_unverified")
     return "The agent finished without a response. Reply in Cabane to try again.";
   if (reason === "session_start_failed")
@@ -12700,11 +12700,11 @@ function turnFailureCopy(reason, runtime) {
   if (reason?.startsWith("seq_floor_unavailable:"))
     return "The interrupted turn could not safely continue. Reply in Cabane to continue.";
   if (reason === "result_error:error_max_budget_usd")
-    return "The agent reached its spending limit. Check its budget settings before trying again in Cabane.";
+    return "The agent reached their spending limit. Check their budget settings before trying again in Cabane.";
   if (reason === "result_error:error_max_structured_output_retries")
     return "The agent could not produce the requested response format. Check the requested format, then reply in Cabane to try again.";
   if (reason?.startsWith("result_error:"))
-    return "The agent stopped with an error while running. Check its transcript, then reply in Cabane to try again.";
+    return "The agent stopped with an error while running. Check their transcript, then reply in Cabane to try again.";
   if (reason === "setup_failed")
     return "The turn could not be set up. Check the agent setup on this device, then reply in Cabane to try again.";
   if (reason === "unexpected_role")
@@ -14586,7 +14586,7 @@ var CompanionSupervisor = class {
       if (!credential) {
         this.log.error(
           { workspaceId, agentId: it.agentId, username: it.agentUsername },
-          `Cannot run ${it.agentDisplayName}: its sign-in to Cabane is missing and Cabane didn't send a new one. Asking again on the next check.`
+          `Cannot run ${it.agentDisplayName}: their sign-in to Cabane is missing and Cabane didn't send a new one. Asking again on the next check.`
         );
         this.removeAgent(wr, it.agentId);
         this.hub.setAgent(workspaceId, {

package/dist/runtime.js

@@ -1756,7 +1756,7 @@ function turnFailureCopy(reason, runtime) {
   if (reason === "timeout_total")
     return "The turn ran longer than expected and was stopped. Reply in Cabane to continue.";
   if (reason === "result_error:error_max_turns")
-    return "The agent reached its step limit. Reply in Cabane to continue.";
+    return "The agent reached their step limit. Reply in Cabane to continue.";
   if (reason === "no_result" || reason === "empty_result" || reason === "empty_result_unverified")
     return "The agent finished without a response. Reply in Cabane to try again.";
   if (reason === "session_start_failed")
@@ -1770,11 +1770,11 @@ function turnFailureCopy(reason, runtime) {
   if (reason?.startsWith("seq_floor_unavailable:"))
     return "The interrupted turn could not safely continue. Reply in Cabane to continue.";
   if (reason === "result_error:error_max_budget_usd")
-    return "The agent reached its spending limit. Check its budget settings before trying again in Cabane.";
+    return "The agent reached their spending limit. Check their budget settings before trying again in Cabane.";
   if (reason === "result_error:error_max_structured_output_retries")
     return "The agent could not produce the requested response format. Check the requested format, then reply in Cabane to try again.";
   if (reason?.startsWith("result_error:"))
-    return "The agent stopped with an error while running. Check its transcript, then reply in Cabane to try again.";
+    return "The agent stopped with an error while running. Check their transcript, then reply in Cabane to try again.";
   if (reason === "setup_failed")
     return "The turn could not be set up. Check the agent setup on this device, then reply in Cabane to try again.";
   if (reason === "unexpected_role")
@@ -3079,7 +3079,7 @@ var TURN_CONTROL_TOOLS = [
   "withdraw_ask"
 ];
 var TURN_CONTROL_TOOL_DESCRIPTIONS = {
-  ask: `Ask a person a structured question you need answered to continue, then end your turn; don't wait for the reply. \`targetUserId\` is a person on your roster, by their \`@handle\`; to bring in an agent, name it with \`@handle\` in your reply.
+  ask: `Ask a person a structured question you need answered to continue, then end your turn; don't wait for the reply. \`targetUserId\` is a person on your roster, by their \`@handle\`; to bring in an agent, name them with \`@handle\` in your reply.
 - Form: one question (\`headline\`, optional \`question\` body, \`options\`) or \`questions\`, 1\u2013${MAX_ASK_ITEMS} \`{ headline, body?, options? }\`; never both. Decisions that arise together go in ONE call, in the order you want them answered. A second \`ask\` is a separate item, never an edit; raise one only for a genuinely new question.
 - \`headline\`: one sentence ending in \`?\`. Body (\`question\` or each item's \`body\`): a sentence or two of framing, not a report; inline Markdown only, no lists or headings. \`options\`: ${MIN_ASK_OPTIONS}\u2013${MAX_ASK_OPTIONS} one-click answers (always for a yes/no go-ahead), each a distinct course of action you can carry out without asking back.
 - Voice, in headline, body and options: "I" for yourself, never your own name; name everyone else who decides or acts, as on your roster. Never "you", "your", "we", "us", "our" or their contractions, nor "he", "she" or "they" for a named actor. Each option reads correctly alone. A quotation stays accurate: say who said it.
@@ -10196,7 +10196,7 @@ var CompanionSupervisor = class {
       if (!credential) {
         this.log.error(
           { workspaceId, agentId: it.agentId, username: it.agentUsername },
-          `Cannot run ${it.agentDisplayName}: its sign-in to Cabane is missing and Cabane didn't send a new one. Asking again on the next check.`
+          `Cannot run ${it.agentDisplayName}: their sign-in to Cabane is missing and Cabane didn't send a new one. Asking again on the next check.`
         );
         this.removeAgent(wr, it.agentId);
         this.hub.setAgent(workspaceId, {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@cabane/companion",
-  "version": "0.6.197",
+  "version": "0.6.199",
   "type": "module",
   "description": "The Cabane Companion (headless): connect an agent on your machine to your Cabane workspace as a responder — drive work against your own codebase, files, and MCP servers without putting any of it in Cabane.",
   "license": "UNLICENSED",