feat: allow up to 20 guidance prompts to wait for the next step

10 was a conservative guess rather than a measured limit. The queue drains at the next
request boundary, so this only caps how many instructions can pile up faster than one
step of the turn (a long tool run or a long streaming answer).

Raising it to 20 keeps that burst covered while still bounding the cost: every queued
prompt travels with each remaining request of the turn, so an unbounded queue could
inflate the context (and any attached images) without limit.

- the queue is not dropped or reordered; it still drains in FIFO order before every
  LLM call of the turn
- when the limit is reached the prompt is refused but stays in the input box, so
  nothing the user typed is lost

Refs lessweb#113, lessweb#117

Author: mulyawansentosa

README-en.md

@@ -102,7 +102,7 @@ turn. The model therefore reads it together with the work it already did and can
 supersede the earlier instructions. If the model is in the middle of writing an answer, press
 `Ctrl+Enter` to cut that answer short so the prompt is read immediately - its text stays in the
 conversation. Set `steerMode: "interrupt"` to make `Ctrl+Enter` the default for `Enter` as well.
-Up to 10 messages can wait; press `Backspace` on an empty prompt to drop the last one. Running
+Up to 20 messages can wait; press `Backspace` on an empty prompt to drop the last one. Running
 commands are never interrupted, `Esc` still interrupts the turn immediately, and slash commands
 still wait for the turn to finish.
 

README.md

@@ -98,7 +98,7 @@ Skills 会按以下优先级扫描：
 AI 正在回复时，`Enter` 不再被拒绝，也不必等到本轮结束：消息会作为补充指引，在本轮下一次
 LLM 调用前作为 user 消息追加到对话中。模型因此能读到它，并结合已完成的工作修改甚至推翻之前的
 指令。如果模型正在输出回答，按 `Ctrl+Enter` 可以截断该回答、立即读取新指令（已生成内容仍保留在
-对话里）；把 `steerMode` 设为 `"interrupt"` 则让普通 `Enter` 也具备同样行为。最多可排队 10 条；
+对话里）；把 `steerMode` 设为 `"interrupt"` 则让普通 `Enter` 也具备同样行为。最多可排队 20 条；
 输入框为空时按 `Backspace` 可移除最后一条。正在执行的命令不会被中断，按 `Esc` 仍会立即中断本轮，
 斜杠命令仍需等待本轮结束。
 

packages/core/src/session.ts

@@ -374,8 +374,15 @@ export type SupplementaryPrompt = {
   createTime: string;
 };
 
-/** Maximum number of prompts that may wait for the next LLM call of a turn. */
-export const MAX_SUPPLEMENTARY_PROMPTS = 10;
+/**
+ * Maximum number of prompts that may wait for the next LLM call of a turn.
+ *
+ * The queue drains at the next request boundary, so this only caps how many
+ * instructions can pile up faster than one step of the turn. Every queued prompt
+ * travels with each remaining request of that turn, so the limit keeps a burst
+ * from inflating the context (and any attached images) without bound.
+ */
+export const MAX_SUPPLEMENTARY_PROMPTS = 20;
 
 export type SessionManagerOptions = {
   projectRoot: string;