@hanskristoffer/agentpool 0.22.0 → 0.24.0

package/README.md

@@ -47,7 +47,7 @@ agentgate service install           # launchd (macOS) or systemd --user (Linux)
 
 You can do every setup step with the CLI. To use the native app, build it with `bun run build:desktop` and open **Agentgate.app**. It connects to `http://127.0.0.1:7878` by default. On a fresh machine, **Set up this machine** installs the bundled CLI in `~/.config/agentgate/bin`, initializes the store, and starts a launchd service. Add that folder to your shell PATH to use the CLI. Closing or removing the app leaves that service and its CLI running.
 
-The app has Overview, Accounts, MCP servers, Skills, Projects, Machines, T3 Code, and Settings screens; request activity is a tab in Settings. **Settings → Configure coding tools** generates the Claude/Codex folders and T3 settings. The app's service and coding-tool controls apply only to its configured local daemon; remote setup and services use the CLI on that machine.
+The app has Overview, Accounts, MCP servers, Skills, Projects, Machines, T3 Code, and Settings screens; request activity is a tab in Settings. **Settings → Configure coding tools** generates the Claude/Codex folders and T3 settings. The app's service and coding-tool controls apply only to its configured local daemon; remote setup and services use the CLI on that machine. The one exception is routing: Machines shows whether each machine's own Claude Code and Codex go through agentgate, and its row menu turns that on or off there.
 
 ## A second machine (e.g. an always-on server)
 
@@ -74,6 +74,15 @@ An Infisical login expires after a few days and can't be renewed, so you have to
 
 To control the server from the native app over Tailscale, open the connection settings at the bottom of the sidebar, enter `http://srv.<tailnet>.ts.net:7878`, and paste the token printed by `agentgate admin-token` on the server. Remote management uses bearer authentication over Tailscale. The relay carries sync only, not remote management.
 
+## Machine dashboard
+
+Each machine can show its own health: CPU, load, memory, disk, time spent waiting for CPU, memory or disk, disk and network traffic, the agents and CI runners it is running, Docker's disk use and containers, the programs using the most CPU, and the last hour as charts. In **Machines**, turn on **Dashboard** for a machine, then **Open dashboard** opens it in your browser. The page is `/health` on the machine's address, and `agentgate dashboard on|off [--node <node>]` switches it from the CLI.
+
+- It runs on Linux machines only, since it reads `/proc`. On a Mac, opening it only says so.
+- Anyone on your tailnet can open it without signing in, as they could the server health page before. It shows program names and resource use, never command lines, credentials or accounts. Turn it off on machines you don't want it on.
+- Turning it on waits until every machine runs this release, because older releases drop the setting when they save a machine's record.
+- The history is kept in memory for the last hour, so a restart starts the charts over.
+
 ## T3 Code
 
 `agentgate setup` writes a Claude config dir (whose `settings.json` points Claude Code at the daemon) and a Codex home. In T3 Code, add one provider instance per provider:
@@ -89,7 +98,7 @@ One instance per provider covers every account; the daemon switches eligible acc
 
 ### Or: your normal Claude login
 
-`agentgate setup --primary` adds `ANTHROPIC_BASE_URL` to `~/.claude/settings.json` and the agentgate MCP server to `~/.claude.json`, so the Claude Code you already use (and T3's default Claude instance) goes through agentgate while keeping its own login. Where Claude Code has never been signed in, such as on a new server, it also sets a placeholder `ANTHROPIC_AUTH_TOKEN`, because Claude Code won't start without a login; agentgate signs each request with a pooled account either way. Claude Desktop is not affected: it sets its own API address; see [Claude Desktop](docs/user/claude-desktop.md). For only the MCP servers, use `agentgate setup --mcp`. That, and `setup --primary`, also write `~/.cursor/mcp.json`, so Cursor and the Cursor CLI get the same servers. Start a new session to pick them up. Model requests use the pool; other requests keep your login, and your login is also the last resort when every pooled account is exhausted. Undo with `agentgate setup --primary off` (the previous base URL is restored, and later settings edits are preserved). It does the same for `~/.codex/config.toml` (or `$CODEX_HOME`): an `agentgate` model provider with `requires_openai_auth = true`, so Codex, the Codex app and T3's default Codex instance send their own ChatGPT login along as the last resort. Only do this with `agentgate service install`, or Claude Code and Codex can't reach their providers while the daemon is down.
+`agentgate setup --primary` adds `ANTHROPIC_BASE_URL` to `~/.claude/settings.json` and the agentgate MCP server to `~/.claude.json`, so the Claude Code you already use (and T3's default Claude instance) goes through agentgate while keeping its own login. Where Claude Code has never been signed in, such as on a new server, it also sets a placeholder `ANTHROPIC_AUTH_TOKEN`, because Claude Code won't start without a login; agentgate signs each request with a pooled account either way. Claude Desktop is not affected: it sets its own API address; see [Claude Desktop](docs/user/claude-desktop.md). For only the MCP servers, use `agentgate setup --mcp`. That, and `setup --primary`, also write `~/.cursor/mcp.json`, so Cursor and the Cursor CLI get the same servers. Start a new session to pick them up. Model requests use the pool; other requests keep your login, and your login is also the last resort when every pooled account is exhausted. Undo with `agentgate setup --primary off` (the previous base URL is restored, and later settings edits are preserved). It does the same for `~/.codex/config.toml` (or `$CODEX_HOME`): an `agentgate` model provider with `requires_openai_auth = true`, so Codex, the Codex app and T3's default Codex instance send their own ChatGPT login along as the last resort. Where Codex has never been signed in, it writes a placeholder API key to `~/.codex/auth.json`, so `codex login status` and T3 Code see Codex as signed in; undo removes it again. Add `--node <name>` to switch a paired machine from the one you are on. Only do this with `agentgate service install`, or Claude Code and Codex can't reach their providers while the daemon is down.
 
 ### Hand threads between machines
 
@@ -213,7 +222,7 @@ Unlike sync, this traffic is readable by the relay (your logins stay on your mac
 
 ## Commands
 
-Run `agentgate --help`. Useful ones: `status`, `accounts`, `accounts exhaust <id> [min]` (pretend an account hit its limit, for testing), `nodes`, `unpair <node>`, `relay status|reconcile|rotate|leave`, `export [--no-secrets]`, `import-backup <file>`, `service logs`.
+Run `agentgate --help`. Useful ones: `status`, `accounts`, `accounts exhaust <id> [min]` (pretend an account hit its limit, for testing), `nodes`, `dashboard on|off`, `unpair <node>`, `relay status|reconcile|rotate|leave`, `export [--no-secrets]`, `import-backup <file>`, `service logs`.
 
 ## Things to know
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@hanskristoffer/agentpool",
-  "version": "0.22.0",
+  "version": "0.24.0",
   "license": "UNLICENSED",
   "repository": {
     "type": "git",
@@ -22,9 +22,9 @@
     "agentgate": "bin/agentpool.js"
   },
   "optionalDependencies": {
-    "agentpool-darwin-arm64": "0.22.0",
-    "agentpool-darwin-x64": "0.22.0",
-    "agentpool-linux-x64": "0.22.0",
-    "agentpool-linux-arm64": "0.22.0"
+    "agentpool-darwin-arm64": "0.24.0",
+    "agentpool-darwin-x64": "0.24.0",
+    "agentpool-linux-x64": "0.24.0",
+    "agentpool-linux-arm64": "0.24.0"
   }
 }