Story: Fix product-backlog bucket names in scripts and captures

The product-backlog directories were renamed from near/ / far/ to next/ / deferred/, but the tooling was not updated in sync. regenerate_backlog_indexes.py still writes near.org and far.org, compass help text still says near|far, and every capture in next/ carries a stale :near: filetag. This story fixes all tooling and capture metadata to match the renamed directories.

• regenerate_backlog_indexes.py writes to next.org and deferred.org, not near.org / far.org.
• near.org, far.org, and the empty near/ and far/ directories are deleted.
• All captures under next/ carry :next: filetag, not :near:.
• All captures under deferred/ carry :deferred: filetag, not :far:.
• compass.py help text uses next|deferred|discarded.
• doc_generate.py comment uses next | deferred.
• Running regenerate_backlog_indexes.py produces no diff against committed next.org / deferred.org.

• near.org and far.org are deleted outright — they are stale duplicates of next.org / deferred.org (same UUID, wrong filename).
• The single file in near/ (extend_validate_docs_frontmatter.org) is moved to next/ as part of the rename cleanup.
• This story exists because Per-sprint capture workflow renamed the directories (near/ → next/, far/ → deferred/) and created the new index files, but did not update regenerate_backlog_indexes.py, compass.py, or the :near: / :far: filetags in existing captures. The script was left pointing at the old filenames.

• The inbox/, discarded/ buckets — names are already current.
• Sprint story/task docs that contain "near" or "far" in prose.