fix: forget search result memories

[Dhravya]

Fixes the forget tool for IDs returned by v4 memory search while preserving deletion for document IDs returned by list.

Changes

• Use v4 memory forgetting as the primary operation across the canonical and scope-relevant legacy containers.
• Fall back to v3 document deletion only after the memory ID is absent from all relevant v4 containers.
• Preserve useful authorization/server errors without letting inaccessible legacy containers block valid document deletion.
• Mark hybrid chunk-only search results as non-forgettable and omit their unusable chunk IDs.

Testing

• PATH='/home/ubuntu/.bun/bin:'"$PATH" bun test — passed, 6 tests and 0 failures.
• PATH='/home/ubuntu/.bun/bin:'"$PATH" bun run typecheck — passed.
• PATH='/home/ubuntu/.bun/bin:'"$PATH" bun run build — passed.
• node dist/cli.js --help — passed.
• Fake-transport harnesses verified v4-first routing, tag order/deduplication, all-404 v3 fallback, canonical and legacy authorization handling, server-error behavior, scoped container routing, and memory/chunk search formatting — passed.

Fixes #66.

---

Session Details

• Session: View Session
• Requested by: Dhravya Shah (dhravya@supermemory.com)
• Address comments on this PR. Add (aside) to your comment to have me ignore it.

[Dhravya]

• fix: forget search result memories #67 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.