fix(screentracker): make writeStabilize Phase 1 non-fatal when agents don't echo input

lib/httpapi/claude.go

@@ -5,6 +5,11 @@ import (
 	st "github.com/coder/agentapi/lib/screentracker"
 )
 
+// formatPaste wraps message in bracketed paste escape sequences.
+// These sequences start with ESC (\x1b), which TUI selection
+// widgets (e.g. Claude Code's numbered-choice prompt) interpret
+// as "cancel". For selection prompts, callers should use
+// MessageTypeRaw to send raw keystrokes directly instead.
 func formatPaste(message string) []st.MessagePart {
 	return []st.MessagePart{
 		// Bracketed paste mode start sequence

lib/msgfmt/message_box.go

@@ -4,15 +4,22 @@ import (
 	"strings"
 )
 
+// containsHorizontalBorder reports whether the line contains a
+// horizontal border made of box-drawing characters (─ or ╌).
+func containsHorizontalBorder(line string) bool {
+	return strings.Contains(line, "───────────────") ||
+		strings.Contains(line, "╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌")
+}
+
 // Usually something like
-// ───────────────
+// ───────────────  (or ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌)
 // >
-// ───────────────
+// ───────────────  (or ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌)
 // Used by Claude Code, Goose, and Aider.
 func findGreaterThanMessageBox(lines []string) int {
 	for i := len(lines) - 1; i >= max(len(lines)-6, 0); i-- {
 		if strings.Contains(lines[i], ">") {
-			if i > 0 && strings.Contains(lines[i-1], "───────────────") {
+			if i > 0 && containsHorizontalBorder(lines[i-1]) {
 				return i - 1
 			}
 			return i
@@ -22,14 +29,14 @@ func findGreaterThanMessageBox(lines []string) int {
 }
 
 // Usually something like
-// ───────────────
+// ───────────────  (or ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌)
 // |
-// ───────────────
+// ───────────────  (or ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌)
 func findGenericSlimMessageBox(lines []string) int {
 	for i := len(lines) - 3; i >= max(len(lines)-9, 0); i-- {
-		if strings.Contains(lines[i], "───────────────") &&
+		if containsHorizontalBorder(lines[i]) &&
 			(strings.Contains(lines[i+1], "|") || strings.Contains(lines[i+1], "│") || strings.Contains(lines[i+1], "❯")) &&
-			strings.Contains(lines[i+2], "───────────────") {
+			containsHorizontalBorder(lines[i+2]) {
 			return i
 		}
 	}

lib/msgfmt/testdata/initialization/claude/ready/msg_dashed_box.txt

@@ -0,0 +1,8 @@
+  1  function greet() {                                                         
+  2 -  console.log("Hello, World!");                                            
+  2 +  console.log("Hello, Claude!");                                           
+  3  }                                                                          
+ ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌ 
+ > Try "what does this code do?"                                                
+ ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌ 
+  Syntax theme: Monokai Extended (ctrl+t to disable)                            

lib/msgfmt/testdata/initialization/claude/ready/msg_slim_dashed_box.txt

@@ -0,0 +1,8 @@
+╭────────────────────────────────────────────╮
+│ ✻ Welcome to Claude Code!                  │
+│                                            │
+│   /help for help                           │
+╰────────────────────────────────────────────╯
+ ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
+ │ Type your message...
+ ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌

lib/screentracker/pty_conversation.go

@@ -3,6 +3,7 @@ package screentracker
 import (
 	"context"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"log/slog"
 	"os"
@@ -16,6 +17,26 @@ import (
 	"golang.org/x/xerrors"
 )
 
+const (
+	// writeStabilizeEchoTimeout is the timeout for the echo
+	// detection WaitFor loop in writeStabilize Phase 1. The
+	// effective ceiling may be slightly longer because the
+	// stability check inside the condition runs outside
+	// WaitFor's timeout select. Non-echoing agents (e.g. TUI
+	// agents using bracketed paste) will hit this timeout,
+	// which is non-fatal.
+	//
+	// TODO: move to PTYConversationConfig if agents need
+	// different echo detection windows.
+	writeStabilizeEchoTimeout = 2 * time.Second
+
+	// writeStabilizeProcessTimeout is the maximum time to wait
+	// for the screen to change after sending a carriage return.
+	// This detects whether the agent is actually processing the
+	// input.
+	writeStabilizeProcessTimeout = 15 * time.Second
+)
+
 // A screenSnapshot represents a snapshot of the PTY at a specific time.
 type screenSnapshot struct {
 	timestamp time.Time
@@ -411,17 +432,30 @@ func (c *PTYConversation) sendMessage(ctx context.Context, messageParts ...Messa
 	return nil
 }
 
-// writeStabilize writes messageParts to the screen and waits for the screen to stabilize after the message is written.
+// writeStabilize writes messageParts to the PTY and waits for
+// the agent to process them. It operates in two phases:
+//
+// Phase 1 (echo detection): writes the message text and waits
+// for the screen to change and stabilize. This detects agents
+// that echo typed input. If the screen doesn't change within
+// writeStabilizeEchoTimeout, this is non-fatal. Many TUI
+// agents buffer bracketed-paste input without rendering it.
+//
+// Phase 2 (processing detection): writes a carriage return
+// and waits for the screen to change, indicating the agent
+// started processing. This phase is fatal on timeout: if the
+// agent doesn't react to Enter, it's unresponsive.
 func (c *PTYConversation) writeStabilize(ctx context.Context, messageParts ...MessagePart) error {
 	screenBeforeMessage := c.cfg.AgentIO.ReadScreen()
 	for _, part := range messageParts {
 		if err := part.Do(c.cfg.AgentIO); err != nil {
 			return xerrors.Errorf("failed to write message part: %w", err)
 		}
 	}
-	// wait for the screen to stabilize after the message is written
+	// Phase 1: wait for the screen to stabilize after the
+	// message is written (echo detection).
 	if err := util.WaitFor(ctx, util.WaitTimeout{
-		Timeout:     15 * time.Second,
+		Timeout:     writeStabilizeEchoTimeout,
 		MinInterval: 50 * time.Millisecond,
 		InitialWait: true,
 		Clock:       c.cfg.Clock,
@@ -441,14 +475,26 @@ func (c *PTYConversation) writeStabilize(ctx context.Context, messageParts ...Me
 		}
 		return false, nil
 	}); err != nil {
-		return xerrors.Errorf("failed to wait for screen to stabilize: %w", err)
-	}
-
-	// wait for the screen to change after the carriage return is written
+		if !errors.Is(err, util.WaitTimedOut) {
+			// Context cancellation or condition errors are fatal.
+			return xerrors.Errorf("failed to wait for screen to stabilize: %w", err)
+		}
+		// Phase 1 timeout is non-fatal: the agent may not echo
+		// input (e.g. TUI agents buffer bracketed-paste content
+		// internally). Proceed to Phase 2 to send the carriage
+		// return.
+		c.cfg.Logger.Info(
+			"echo detection timed out, sending carriage return",
+			"timeout", writeStabilizeEchoTimeout,
+		)
+	}
+
+	// Phase 2: wait for the screen to change after the
+	// carriage return is written (processing detection).
 	screenBeforeCarriageReturn := c.cfg.AgentIO.ReadScreen()
 	lastCarriageReturnTime := time.Time{}
 	if err := util.WaitFor(ctx, util.WaitTimeout{
-		Timeout:     15 * time.Second,
+		Timeout:     writeStabilizeProcessTimeout,
 		MinInterval: 25 * time.Millisecond,
 		Clock:       c.cfg.Clock,
 	}, func() (bool, error) {
@@ -527,6 +573,14 @@ func (c *PTYConversation) statusLocked() ConversationStatus {
 		return ConversationStatusChanging
 	}
 
+	// The send loop gates stableSignal on initialPromptReady.
+	// Report "changing" until readiness is detected so that Send()
+	// rejects with ErrMessageValidationChanging instead of blocking
+	// indefinitely on a stableSignal that will never fire.
+	if !c.initialPromptReady {
+		return ConversationStatusChanging
+	}
+
 	// Handle initial prompt readiness: report "changing" until the queue is drained
 	// to avoid the status flipping "changing" -> "stable" -> "changing"
 	if len(c.outboundQueue) > 0 || c.sendingMessage {