Teach parent agents to clean up the processes they spawn

Add a `lifecycle` help topic spelling out that the caller owns the
processes it spawns and should `close_process` when a sub-agent or
spawned child is no longer needed. The `spawn_agent` and `spawn_process`
descriptions advertised via `tools/list` now restate the same duty
inline (with a pointer to `help('lifecycle')`), so vendor TUIs see the
expectation at the moment they reach for the tool. The `spawning` topic
and `topics` index cross-reference the new content.

Bundles two already-staged improvements that fall in the same area:
- OnChildSpawned primes the snapshot-replay budget for new panes so
  diff-based vendor TUIs come up clean without a manual Ctrl+W/Ctrl+S
  refresh.
- TODO drops the three items now actioned (prompt-injection preface,
  agent cleanup duty, opencode→claude view corruption) and keeps the
  unicode `<?>` entry with the investigation notes.

internal/app/host.go

@@ -140,7 +140,8 @@ func (h *toolHost) SpawnAgent(callerID string, args mcp.SpawnAgentArgs) (mcp.Pro
 	if display == "" {
 		display = args.Agent
 	}
-	c, err := h.launcher.LaunchAgent(p, display, args.AgentInstructions, callerID)
+	prompt := wrapSubAgentPrompt(args.AgentInstructions, h.sess.FindChild(callerID) != nil)
+	c, err := h.launcher.LaunchAgent(p, display, prompt, callerID)
 	if err != nil {
 		return mcp.ProcessInfo{}, err
 	}
@@ -815,6 +816,25 @@ func (h *toolHost) askForTrust(callerID, presetName, reason string) {
 	h.prompter.promptTrust(callerID, presetName, reason)
 }
 
+// wrapSubAgentPrompt prepends a one-line orientation block to the
+// caller-supplied agent_instructions. patterm injects nothing on its
+// own (SPEC §7), but vendor TUIs that learn their role purely from
+// their first turn need to be told they're a sub-agent — otherwise
+// they finish without reporting back to the parent or cleaning up
+// processes/scratchpads they spawned. The block is single-line on
+// purpose: writeInput splits on CR/LF, so any embedded newline would
+// submit prematurely.
+func wrapSubAgentPrompt(instructions string, hasParent bool) string {
+	if !hasParent {
+		return instructions
+	}
+	if instructions == "" {
+		return ""
+	}
+	const preface = "[system: you are a patterm sub-agent. When your work is done, call send_message to your parent (use whoami to get parent_process_id) with a summary, and close_process / scratchpad cleanup anything you created. See help('conventions').] "
+	return preface + instructions
+}
+
 // applyChromeTrim deletes lines matching any of the given regexes.
 // SPEC §10 chrome_trim_hints.
 func applyChromeTrim(txt string, hints []string) string {
@@ -894,7 +914,7 @@ func helpFor(topic string) mcp.HelpResponse {
 	case "", "topics":
 		return mcp.HelpResponse{
 			Topic: "topics",
-			Content: "Available topics: spawning, inspection, io, coordination, " +
+			Content: "Available topics: spawning, lifecycle, inspection, io, coordination, " +
 				"scratchpads, timers, readiness, permissions, conventions, topics. " +
 				"Call help(topic) for guidance. Call whoami for your role and the " +
 				"complete tool list available to you.",
@@ -902,8 +922,14 @@ func helpFor(topic string) mcp.HelpResponse {
 	case "spawning":
 		return mcp.HelpResponse{
 			Topic:        "spawning",
-			Content:      "spawn_agent launches another vendor LLM CLI as a sub-agent (orchestrator only). spawn_process(kind: command, preset: …) starts a stored command; spawn_process(kind: terminal) opens a shell. Command presets need trust the first time — you'll get needs_trust until the human accepts.",
-			RelatedTools: []string{"spawn_agent", "spawn_process", "start_process", "restart_process"},
+			Content:      "spawn_agent launches another vendor LLM CLI as a sub-agent (orchestrator only). spawn_process(kind: command, preset: …) starts a stored command; spawn_process(kind: terminal) opens a shell. Command presets need trust the first time — you'll get needs_trust until the human accepts. Whatever you spawn is yours to clean up — see help('lifecycle').",
+			RelatedTools: []string{"spawn_agent", "spawn_process", "start_process", "restart_process", "close_process"},
+		}
+	case "lifecycle":
+		return mcp.HelpResponse{
+			Topic: "lifecycle",
+			Content: "You own the processes you spawn. When a sub-agent has finished its task (it reports back via send_message, or you've collected what you need from it) call close_process on its process_id to remove the entry and tear down the PTY. Same goes for spawn_process children: command/terminal panes you started are not auto-reclaimed when their work completes. close_process is the normal cleanup path; stop_process(signal) is for sending a signal without removing the entry; start_process re-attaches an exited command preset. Leaving idle sub-agents around wastes vendor tokens and clutters the host — close them as soon as you're done. Sub-agents themselves are reminded (via the [system: …] preface on their first prompt) to clean up anything they created before reporting done.",
+			RelatedTools: []string{"close_process", "stop_process", "start_process", "list_processes", "get_process_status"},
 		}
 	case "inspection":
 		return mcp.HelpResponse{