feat: full Copilot SDK feature parity — custom agents, hooks, image input, skills, steering, session persistence

Backend (sidecar):
- Extended ProtocolModels with DTOs for custom agents, hooks, skills,
  infinite sessions, session lifecycle, and 9 new event types
- Added CopilotManagedSessionIds for stable SDK session ID mapping
- Added CopilotSessionManager/ICopilotSessionManager for session lifecycle
- Added CopilotSessionHooks for hook registration
- Added CopilotMessageOptionsMetadata for mid-turn steering
- Extended CopilotAgentBundle to wire custom agents, hooks, skills,
  infinite sessions, and stable session IDs
- Extended CopilotTurnExecutionState to project 13 new SDK event types
- Widened ITurnWorkflowRunner callback to accept SidecarEventDto
- Added list/delete/disconnect session commands to SidecarProtocolHost
- Added AryxCopilotAgentMessageOptionsTests (14 new tests, 142 total)

Frontend (renderer + main + shared):
- Added ChatMessageAttachment type and helpers (attachment.ts)
- Extended sidecar contracts with MessageMode, 3 new command types,
  9 new event types, and agent/session config DTOs
- Extended SessionEventRecord with 6 new event kinds and ~20 fields
- Added PatternAgentCopilotConfig to pattern domain
- Added attachments support to ChatMessageRecord
- Updated sidecar client with session lifecycle methods and
  turn-scoped event routing via onTurnScopedEvent callback
- Updated main process: handleTurnScopedEvent(), deleteSession(),
  steering bypass for mid-turn messages, attachment passthrough
- Added deleteSession IPC handler and preload binding
- Added TurnEventLog state tracker with format/apply/prune helpers
- ChatPane: always-enabled composer, steering indicator, attachment
  picker with preview, image thumbnails in message history,
  context-usage bar, amber steer mode for send button
- ActivityPanel: turn events section with sub-agent, hook, skill,
  and compaction event rendering
- Sidebar: delete session action in context menu
- App.tsx: wired sessionUsage, turnEventLogs, and deleteSession

Documentation:
- AGENTS.md: added glob safety rule for node_modules
- README.md: added steering, image input, and richer observability
- ARCHITECTURE.md: added turn-scoped events, steering, and attachments
- Website: added steering and image input feature cards, updated
  live visibility and session cards

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

README.md

@@ -17,8 +17,10 @@ It works especially well when you want AI help that stays grounded in an actual
 - **Start fast** with a scratchpad conversation for quick questions and ad-hoc work.
 - **Work against real projects** by attaching local folders and letting Aryx stay aware of repository context.
 - **Go beyond one assistant** with orchestration patterns such as single-agent, sequential, concurrent, handoff, and group-chat flows.
-- **See what is happening** with live activity for each agent while a run is in progress.
-- **Stay organized** with persistent sessions you can rename, pin, archive, and return to later.
+- **See what is happening** with live activity for each agent while a run is in progress, including sub-agent delegations, hook lifecycle, skill invocations, and context compaction.
+- **Stay organized** with persistent sessions you can rename, pin, archive, delete, and return to later.
+- **Steer while agents work** by sending follow-up messages mid-turn — the agent receives your input immediately.
+- **Attach images** to any message for visual context the model can reason about.
 - **Tune how you work** by choosing models and reusing saved patterns that fit different tasks.
 
 ## What you can do in the app
@@ -51,11 +53,19 @@ Patterns now require tool-call approval by default. They can also store default
 
 ### Watch runs as they happen
 
-You can follow agent activity while a session is running, which makes longer or more complex workflows easier to trust and understand.
+You can follow agent activity while a session is running, which makes longer or more complex workflows easier to trust and understand. The activity panel shows sub-agent delegations, skill invocations, hook lifecycle events, and context compaction in real time. A context-usage bar below the composer shows how much of the model's context window the current session occupies.
+
+### Steer agents mid-turn
+
+While an agent is working, you can type a follow-up message that is delivered immediately into the current turn. This lets you redirect, refine, or add context without waiting for the turn to finish. The composer shows an amber "steering" indicator when a message will be injected into an active run.
+
+### Attach images
+
+You can attach images (JPEG, PNG, GIF, WebP) to any message using the clip button, drag-and-drop, or paste from clipboard. Image attachments are sent as base64-encoded blobs so the model can reason about visual content alongside your text.
 
 ### Keep important work around
 
-Sessions are persistent, so you can return to ongoing work instead of starting from scratch every time. You can also rename, pin, archive, and duplicate sessions as your workspace grows.
+Sessions are persistent, so you can return to ongoing work instead of starting from scratch every time. You can also rename, pin, archive, delete, and duplicate sessions as your workspace grows.
 
 ## Before you start