AskHamiz: upload images and documents directly in chat

[shreekardittakavi]

Spec: Doc — "for example the user may upload images/documents directly to the chatbot rather than manually do it. Ultimate laziness agentic chatbot."

Scope

• Attach images and PDFs directly in the chat composer (drag-drop, paste, and file picker).
• Route uploads through the existing notes/upload pipeline so parsing, topic splitting, and embedding behave identically to the Notes page — do not fork a second extraction path.
• The agent should infer intent from the message: "add this to my calc notes" links the result to that class; a bare syllabus should be offered to the syllabus parser instead.
• Show the resulting note inline in the conversation with a link to open it in the editor.
• Enforce the same mime allowlist and size limits as the existing upload route.

Depends on

• AskHamiz agentic tool surface
• /api/chat session guard

Acceptance criteria

• A user can drop a photo of handwritten notes into chat and get a saved, parsed note
• Class association is inferred from the message or asked for when ambiguous
• Rejected file types fail with a clear message, not a silent no-op
• Upload progress and failure states are visible (AGENTS.md §1)