feat(autostart): self-perpetuating work loop — stay reachable

The recurring "agent doesn't read messages in realtime / isn't triggered for the
next task" problem: agents weren't STAYING in the agenthub_work loop, so they went
dormant after each task. Two reinforcements:

- agenthub_work now ends every response with a LOOP directive: "the instant you
  finish, call agenthub_work again (in the background) so the next task/message
  reaches you in realtime — never end your turn without relaunching it."
- The auto-start hook-context now stresses the BACKGROUND loop + relaunch pattern,
  so a Claude Code session is woken on each event instead of blocking.

Layer 1 (auto-start) gets agents INTO the loop; this keeps them IN it. Honest
limit: turn-based agents still rely on obeying the relaunch nudge — there's no
100% always-on without a supervisor. Bump 0.9.0 -> 0.9.1.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

package.json

@@ -1,6 +1,6 @@
 {
   "name": "agenthub",
-  "version": "0.9.0",
+  "version": "0.9.1",
   "description": "Local coordination layer for AI coding agents",
   "type": "module",
   "main": "./dist/index.js",

src/cli/commands/agentSetup.ts

@@ -22,11 +22,14 @@ import { findProjectRoot } from '../../core/paths.js';
 export function hookContext(opts: { agent: string; role: string }): void {
   const { agent, role } = opts;
   process.stdout.write(
-    `AgentHub auto-start — you are "${agent}" (${role}). Before anything else, run:\n` +
+    `AgentHub auto-start — you are "${agent}" (${role}). Before anything else, enter the work loop:\n` +
       `  agenthub work --agent ${agent} --role ${role}\n` +
-      `(or call the agenthub_work tool with { agent: "${agent}", role: "${role}" }), then LOOP: ` +
-      `work -> implement -> agenthub_task_review -> work. agenthub_work claims your tasks AND ` +
-      `surfaces your messages. Never run agenthub_task_done. Keep looping so you stay reachable.\n`,
+      `(or call the agenthub_work tool with { agent: "${agent}", role: "${role}" }). It BLOCKS until a ` +
+      `task or message addressed to you arrives, then returns it. Handle it, then IMMEDIATELY call ` +
+      `agenthub_work AGAIN — run it in the BACKGROUND so you're woken on the next event without ` +
+      `tying up your turn. The loop is: work -> handle -> agenthub_task_review -> work, forever. ` +
+      `agenthub_work claims your tasks AND delivers your messages. Never run agenthub_task_done. ` +
+      `Staying in this loop is what keeps you reachable — do not end a turn without relaunching agenthub_work.\n`,
   );
 }
 

src/cli/index.ts

@@ -118,7 +118,7 @@ async function runRemote(serverUrl: string, fn: () => Promise<void>): Promise<vo
 export function createProgram(cwd: string): Command {
   const program = new Command('agenthub')
     .description('Local coordination layer for AI coding agents')
-    .version('0.9.0')
+    .version('0.9.1')
     .option('--server <url>', 'AgentHub server URL (env: AGENTHUB_SERVER)');
 
   program

src/mcp/server.ts

@@ -135,11 +135,18 @@ export async function startMcpServer(cwd: string): Promise<void> {
         if (messages.length) return { claimed: null, messages, note: 'No task addressed to you, but you have messages — reply with agenthub_message.' };
         return null;
       };
+      // Self-perpetuating loop: every response reminds the agent to relaunch
+      // agenthub_work, so a finished task/message never leaves it dormant.
+      const LOOP =
+        'STAY REACHABLE: the instant you finish handling this, call agenthub_work again — ' +
+        'run it in the background so the next task or message reaches you in realtime. ' +
+        'Never end your turn without relaunching agenthub_work.';
       const immediate = await findWork();
-      if (immediate) return asText(immediate);
+      if (immediate) return asText({ ...immediate, loop: LOOP });
       if (!remote) return asText(`No open task addressed to ${agent}, and no hub server to wait on.`);
       const result = await waitForTask(serverUrl!, findWork, timeoutSec ?? 300);
-      return asText(result ?? `No task or message for ${agent} within ${timeoutSec ?? 300}s — call agenthub_work again.`);
+      if (result) return asText({ ...result, loop: LOOP });
+      return asText(`No task or message for ${agent} within ${timeoutSec ?? 300}s. ${LOOP}`);
     });
 
   server.tool('agenthub_task_list', 'List tasks, optionally filtered by status and/or role.',