fix: address max-effort review findings

- clamp LLM-derived page indices in _get_text_of_pages and
  get_text_of_pdf_pages_with_labels; dedupe get_text_of_pdf_pages
- guard _normalize_tree and folders/documents iterations against
  explicit nulls in cloud API responses
- coerce cloud OCR page numbers to int before filtering in
  get_page_content
- folder cache: raise on missing folder id instead of caching None;
  stop caching name-not-found so later lookups can succeed
- defang doc_id in agent doc-context prompt
- append api-key hint to 401 errors (request, legacy and streaming
  paths)
- remove stale legacy JSON workspace sample data unreadable by the
  SQLite storage

examples/workspace/_meta.json

@@ -1,9 +0,0 @@
-{
-  "12345678-abcd-4321-abcd-123456789abc": {
-    "type": "pdf",
-    "doc_name": "attention-residuals.pdf",
-    "doc_description": "This document introduces \"Attention Residuals\" (AttnRes) and its scalable variant \"Block AttnRes,\" novel mechanisms for replacing fixed residual accumulation in neural networks with learned, input-dependent depth-wise attention, addressing limitations of standard residual connections while optimizing memory, computation, and scalability for large-scale training and inference.",
-    "page_count": 21,
-    "path": "../documents/attention-residuals.pdf"
-  }
-}