chore(release): 0.39.44

[pitimon]

Summary

Release 0.39.44. Closes #125.

Version bumped in all four lockstep locations by the version hook — package.json, package-lock.json, both TokenTrackerBar targets (MARKETING_VERSION), and TokenTrackerWin.csproj. validate:version-lockstep passes.

15 commits since v0.39.43.

What users get

Change	Why it matters
Projects panel honours the date range (#118)	It read no query parameter. Picking "24h" narrowed every other card while Projects kept showing all-time totals, with nothing on screen saying so.
Per-repo cost (#121)	Project rows now carry a model, so computeRowCost works per repository.
Unattributed sources are named (#118)	Cursor, Copilot, Zed, Goose and Kiro have no per-repo story; omitting them looked identical to "that tool cost nothing here".
Plan-value card (#122)	Answers the question README:32 already promised, labelled list-price-equivalent rather than a saving.
Codex chip survives a 401 (#119)	A stale sign-in produced no windows and no error, so the chip disappeared — worse than never having had one.
sync --compact (#117)	34,924 lines to 5,636 on a real install. Sync recommends it past a threshold and does not act on its own.

Behind the scenes: outbound-privacy validator hardening (#111, #114), avatar proxy per-hop revalidation + port-aware allowlisting + a real download cap (#109, #112), a parser conformance ratchet (#116, #120), scripts/graph out of the product gate (#115), a usage-limits fixture capture tool (#124).

The one thing to watch after this ships

#121 migrates cursors.json on every user's first sync after upgrading, re-keying project buckets from project|source|hour to project|source|model|hour. Without that the old bucket strands, new usage restarts from zero under the model key, and the reader still sees the old row — the usage is counted twice.

Verified read-only against a real 2,015-bucket state:

• every key migrated from three segments to four
• total_tokens preserved exactly: 6,281,653,062 unchanged
• all 2,015 queuedKey markers survived — losing those would re-append every row
• the real file was not touched

That is one machine. A differently-shaped state is the residual risk, and it is the thing to look at first if anything is reported.

Test plan

• ci:local exit 0 — 972 root tests, 302 dashboard
• validate:version-lockstep passes; all four locations read 0.39.44
• npm view @ipv9/tokentracker-cli@0.39.44 → 404, so the version is free
• After merge: tag v0.39.44, run release (macOS + Windows), npm publish
• Verify the published package runs and the DMG reports 0.39.44 — not just that the workflow went green
• Repin the local LaunchAgents