fix: Prevent agents page from scrolling — lock layout to viewport

Body: height 100vh + overflow hidden instead of min-height 100vh,
so the browser never shows a scrollbar on html/body.
AppLayout: h-screen flex column with shrink-0 header and flex-1
min-h-0 overflow-auto main. Pages like initiatives scroll within
main; agents page uses h-full with internal panel scrollers.

docs/agent.md

@@ -143,15 +143,27 @@ Agents can communicate with each other via the `conversations` table, coordinate
 
 ### Prompt Integration
 `INTER_AGENT_COMMUNICATION` constant in `prompts/shared.ts` is appended to all 5 agent mode prompts. It instructs agents to:
-1. Start `cw listen --agent-id <ID> &` as a background process at session start
-2. Handle incoming questions by answering via `cw answer` and restarting the listener
-3. Ask questions to peers via `cw ask --from <ID> --agent-id|--phase-id|--task-id`
-4. Kill the listener before writing `signal.json`
+1. Set up a background listener via temp-file redirect: `cw listen > $CW_LISTEN_FILE &`
+2. Periodically check the temp file for incoming questions between work steps
+3. Answer via `cw answer`, clear the file, restart the listener
+4. Ask questions to peers via `cw ask --from <ID> --agent-id|--task-id|--phase-id`
+5. Kill the listener and clean up the temp file before writing `signal.json`
 
 ### Agent Identity
 `manifest.json` now includes `agentId` and `agentName` fields. The manager passes these from the DB record after agent creation.
 
 ### CLI Commands
-- `cw listen`: Polls `getPendingConversations`, prints first pending as JSON, exits
-- `cw ask`: Creates conversation, polls `getConversation` until answered, prints answer
-- `cw answer`: Calls `answerConversation`, prints confirmation JSON
+
+**`cw listen --agent-id <id> [--timeout <ms>] [--poll-interval <ms>]`**
+- Polls `getPendingConversations`, prints first pending as JSON, exits with code 0
+- `--timeout`: max wait in ms (default 0=forever)
+- `--poll-interval`: polling frequency in ms (default 2000)
+- Output: `{ conversationId, fromAgentId, question, phaseId?, taskId? }`
+
+**`cw ask <question> --from <agentId> --agent-id|--task-id|--phase-id <target> [--timeout <ms>] [--poll-interval <ms>]`**
+- Creates conversation, polls `getConversation` until answered, prints answer text to stdout
+- Target resolution: `--agent-id` (direct), `--task-id` (find agent running task), `--phase-id` (find agent in phase)
+- `--timeout` / `--poll-interval`: same defaults as listen
+
+**`cw answer <answer> --conversation-id <id>`**
+- Calls `answerConversation`, prints `{ conversationId, status: "answered" }`

packages/web/src/index.css

@@ -57,7 +57,8 @@
     @apply bg-background text-foreground;
     margin: 0;
     min-width: 320px;
-    min-height: 100vh;
+    height: 100vh;
+    overflow: hidden;
   }
 }
 

packages/web/src/layouts/AppLayout.tsx

@@ -9,9 +9,9 @@ const navItems = [
 
 export function AppLayout({ children }: { children: React.ReactNode }) {
   return (
-    <div className="min-h-screen bg-background">
+    <div className="flex h-screen flex-col bg-background">
       {/* Header */}
-      <header className="sticky top-0 z-50 border-b bg-background/95 backdrop-blur supports-[backdrop-filter]:bg-background/60">
+      <header className="shrink-0 border-b bg-background/95 backdrop-blur supports-[backdrop-filter]:bg-background/60">
         <div className="mx-auto flex h-14 max-w-7xl items-center justify-between px-6">
           <Link to="/initiatives" className="text-lg font-bold tracking-tight">
             Codewalk District
@@ -35,7 +35,7 @@ export function AppLayout({ children }: { children: React.ReactNode }) {
       </header>
 
       {/* Page content */}
-      <main className="mx-auto max-w-7xl px-6 py-6">
+      <main className="mx-auto flex-1 min-h-0 w-full max-w-7xl overflow-auto px-6 py-6">
         {children}
       </main>
     </div>

packages/web/src/routes/agents.tsx

@@ -89,10 +89,7 @@ function AgentsPage() {
   // Loading state
   if (agentsQuery.isLoading) {
     return (
-      <div
-        style={{ height: "calc(100vh - 7rem)" }}
-        className="flex flex-col gap-4"
-      >
+      <div className="flex h-full flex-col gap-4">
         <div className="flex items-center justify-between shrink-0">
           <div className="flex items-center gap-2">
             <Skeleton className="h-6 w-20" />
@@ -179,10 +176,7 @@ function AgentsPage() {
   ];
 
   return (
-    <div
-      style={{ height: "calc(100vh - 7rem)" }}
-      className="flex flex-col gap-4"
-    >
+    <div className="flex h-full flex-col gap-4">
       {/* Header + Filters */}
       <div className="shrink-0 space-y-3">
         <div className="flex items-center justify-between">

src/agent/prompts/shared.ts

@@ -47,53 +47,80 @@ Use the output as the filename (e.g., \`{id}.md\`).`;
 export const INTER_AGENT_COMMUNICATION = `
 ## Inter-Agent Communication
 
-You are working in a multi-agent parallel environment. Other agents may be working on related tasks simultaneously.
+You are working in a multi-agent parallel environment. Other agents may be working on related tasks simultaneously. You can exchange questions and answers with peer agents via CLI commands.
 
 ### Your Identity
-Read \`.cw/input/manifest.json\` — it contains \`agentId\` and \`agentName\` fields identifying you.
+Read \`.cw/input/manifest.json\` — it contains \`agentId\` and \`agentName\` fields identifying you. You'll need your \`agentId\` for all communication commands.
 
-### Listening for Questions
-At the START of your session, start a background listener:
+### CLI Commands
+
+**\`cw listen\`** — Poll for incoming questions. Prints the first pending question as JSON and exits.
+\`\`\`
+cw listen --agent-id <YOUR_AGENT_ID> [--timeout <ms>] [--poll-interval <ms>]
+\`\`\`
+- \`--agent-id\` (required): Your agent ID
+- \`--timeout\`: Max wait in ms. Default: 0 (wait forever). Use a value like 120000 (2 min) to avoid hanging.
+- \`--poll-interval\`: Polling frequency in ms. Default: 2000
+- Output (JSON): \`{ "conversationId": "...", "fromAgentId": "...", "question": "...", "phaseId": "...", "taskId": "..." }\`
+- Exit code 0 if a question was found, 1 on timeout or error.
+
+**\`cw ask\`** — Ask another agent a question. Blocks until the answer arrives, then prints the answer text to stdout.
+\`\`\`
+cw ask "<question>" --from <YOUR_AGENT_ID> --agent-id <TARGET_AGENT_ID> [--timeout <ms>] [--poll-interval <ms>]
+\`\`\`
+- \`--from\` (required): Your agent ID (the asker)
+- Target (exactly one required):
+  - \`--agent-id <id>\`: Ask a specific agent directly
+  - \`--task-id <id>\`: Ask whichever agent is running that task
+  - \`--phase-id <id>\`: Ask whichever agent is running a task in that phase
+- \`--timeout\`: Max wait in ms. Default: 0 (wait forever). Use 120000+ for safety.
+- \`--poll-interval\`: Polling frequency in ms. Default: 2000
+- Output: The answer text (plain text, not JSON).
+- Exit code 0 if answered, 1 on timeout or error.
+
+**\`cw answer\`** — Answer a pending question.
+\`\`\`
+cw answer "<answer>" --conversation-id <CONVERSATION_ID>
+\`\`\`
+- \`--conversation-id\` (required): The conversation ID from the listen output
+- Output (JSON): \`{ "conversationId": "...", "status": "answered" }\`
+
+### Background Listener Pattern
+
+At the START of your session, set up a background listener that writes to a temp file:
 \`\`\`bash
-cw listen --agent-id <YOUR_AGENT_ID> &
+CW_LISTEN_FILE=$(mktemp /tmp/cw-listen-XXXXXX.txt)
+cw listen --agent-id <YOUR_AGENT_ID> --timeout 120000 > "$CW_LISTEN_FILE" 2>&1 &
 LISTEN_PID=$!
 \`\`\`
 
-When the listener prints JSON to stdout, another agent is asking you a question:
-\`\`\`json
-{ "conversationId": "...", "fromAgentId": "...", "question": "..." }
+Periodically check for incoming questions between your work steps:
+\`\`\`bash
+LISTEN_CONTENT=$(cat "$CW_LISTEN_FILE" 2>/dev/null)
+if [ -n "$LISTEN_CONTENT" ]; then
+  echo "$LISTEN_CONTENT"
+fi
 \`\`\`
 
-Answer it:
+When a question arrives, parse the JSON, answer, then restart the listener:
 \`\`\`bash
+# Parse conversationId and question from the JSON
 cw answer "<your answer>" --conversation-id <conversationId>
-\`\`\`
-
-Then restart the listener:
-\`\`\`bash
-cw listen --agent-id <YOUR_AGENT_ID> &
+# Clear and restart
+> "$CW_LISTEN_FILE"
+cw listen --agent-id <YOUR_AGENT_ID> --timeout 120000 > "$CW_LISTEN_FILE" 2>&1 &
 LISTEN_PID=$!
 \`\`\`
 
-### Asking Questions
-To ask another agent a question (blocks until answered):
-\`\`\`bash
-cw ask "What interface does the user service expose?" --from <YOUR_AGENT_ID> --agent-id <TARGET_AGENT_ID>
-\`\`\`
-
-You can also target by task or phase:
-\`\`\`bash
-cw ask "What port does the API run on?" --from <YOUR_AGENT_ID> --task-id <TASK_ID>
-cw ask "What schema are you using?" --from <YOUR_AGENT_ID> --phase-id <PHASE_ID>
-\`\`\`
-
 ### When to Communicate
-- You need interface/schema/contract info from another agent's work
+- You need interface, schema, or API contract info from another agent's work
 - You're about to modify a shared resource and want to coordinate
 - You have a dependency on work another agent is doing
+- Do NOT ask questions that you can answer by reading the codebase yourself
 
 ### Cleanup
 Before writing \`.cw/output/signal.json\`, kill your listener:
 \`\`\`bash
 kill $LISTEN_PID 2>/dev/null
+rm -f "$CW_LISTEN_FILE"
 \`\`\``;

src/preview/manager.test.ts

@@ -73,6 +73,8 @@ function createMockEventBus(): EventBus & { emitted: DomainEvent[] } {
       emitted.push(event);
     }),
     on: vi.fn(),
+    off: vi.fn(),
+    once: vi.fn(),
   };
 }