feat(compound-engineering): live agent output, steering, and a real Q&A surface

Users can now watch everything the agent does while a CE stage works, steer
it mid-stage, and read the whole conversation as a proper chat surface.

Live output:
- New host capability: CreateInteractiveAiSessionOptions.onProgress — the
  engine adapter streams thinking/text deltas + tool start/end markers from
  the pi agent hooks (any plugin can use this).
- Orchestrator buffers per-session live activity (merged deltas, discrete
  tool lines, capped), emits throttled progress events over SSE, and
  GET /sessions/:id attaches it as liveActivity for the polling fallback.
- Routes detach turn execution: start/answer/resume return immediately
  (status active) and clients converge via push/poll — the turn is watchable
  instead of hidden inside a blocking POST.
- Turn timeout is now INACTIVITY-based: an actively-working long turn is
  never killed; a quiet one interrupts with its working trace preserved.
- On settle the trace persists into history as a condensed record.

Steering:
- Stage protocol: responses may be a direct answer, {value, comment}
  (answer + guidance), or {feedback} (guidance without answering); the
  system prompt instructs agents to treat steering as first-class input.
- CeFlow: guidance textarea alongside selectable questions — attach to the
  clicked answer, or "Send guidance" on its own.

Q&A UI:
- Transcript no longer hides control records: past questions/answers render
  as chat bubbles (option ids → labels), steering turns marked, working
  traces as collapsible "Agent work" blocks, completion marker.
- Live working pane (pulse + streaming thinking/tool lines) while a turn runs.

Tests: 130 plugin tests green (14 new: live buffer/flush ordering, inactivity
watchdog survives active work, detached convergence, steering payload shapes,
transcript rendering, live pane). Engine seam tests green; plugin/core/
engine/dashboard tsc clean. Core full suite OOMs locally (known orchestrator-
shell issue) — covered by CI shards.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

packages/core/src/index.ts

@@ -528,6 +528,7 @@ export type {
   AiSessionResult,
   CreateAiSessionFactory,
   CreateInteractiveAiSessionOptions,
+  InteractiveAiSessionProgressEvent,
   InteractiveAiSessionEvent,
   InteractiveAiSession,
   CreateInteractiveAiSessionResult,

packages/core/src/plugin-types.ts

@@ -164,8 +164,29 @@ export interface CreateInteractiveAiSessionOptions {
    * `requestedSkillNames` are actually discoverable in the live session.
    */
   additionalSkillPaths?: string[];
+  /**
+   * Live progress callback, invoked WHILE a turn runs (the pull-based
+   * `nextEvent()` only resolves once the turn settles). Receives streaming
+   * thinking/text deltas and tool start/end markers so a caller can surface
+   * the agent's work in real time. Optional; ignored by factories that cannot
+   * stream. Must not throw — implementations should swallow callback errors.
+   */
+  onProgress?: (event: InteractiveAiSessionProgressEvent) => void;
 }
 
+/**
+ * A live progress event emitted mid-turn via
+ * {@link CreateInteractiveAiSessionOptions.onProgress}.
+ *
+ * - `thinking` / `text`: an incremental output DELTA (not a snapshot) — the
+ *   consumer accumulates.
+ * - `tool`: a discrete tool execution start/end marker.
+ */
+export type InteractiveAiSessionProgressEvent =
+  | { type: "thinking"; delta: string }
+  | { type: "text"; delta: string }
+  | { type: "tool"; name: string; phase: "start" | "end"; isError?: boolean };
+
 /**
  * A single event pulled from an interactive AI session.
  *