v0.7.0 — Thinking depth badge

Highlights

[Opus 4.8 · XHigh] █▊░░░░░░░░ 18% │ 5h:1% (4.5h) │ 7d:35% (4.1d) │ Fable:51%

• Thinking depth replaces the plan-tier badge. The [model · X] slot now shows the model's current reasoning effort — Low / Medium / High / XHigh / Max.
• Read straight from the transcript (the latest turn's effort) — zero network, works on every backend (Claude, DeepSeek, GLM, MiniMax). The badge hides itself when no effort is present.
• The top-model weekly gauge is unchanged.

Upgrade

/plugin marketplace update cc-hud
/reload-plugins

Other changes

• transcript.ts: a single tail read now yields active subagents and the current thinking depth
• Removed the ~/.claude.json plan-tier reading — claude.ts keeps only the OAuth top-model gauge
• 142 tests (added transcript effort coverage); README + animated preview synced