# Changelog

All notable changes to RigDoctor are recorded here. Format follows
[Keep a Changelog](https://keepachangelog.com/); versioning is SemVer-style
(`MAJOR.MINOR.PATCH`, pre-1.0). `__version__` and `pyproject.toml` must match the git
release tag (so the auto-updater, D18, can compare versions).

## [0.38.0] - 2026-05-22
### Added
- **PCIe link in the Inventory.** Each NVMe drive now shows its negotiated PCIe link next to the
  model — e.g. `Samsung SSD 980 PRO 1TB (931.5G) · PCIe Gen4 x4` — read from sysfs
  (`current/max_link_speed` + width). If a drive negotiates below its capability (a slower M.2
  slot, lane-sharing, or a downtrain) it's flagged: `PCIe Gen3 x4 (capable of Gen4 x4)`. So you
  can confirm a Gen4 SSD is actually in a Gen4 slot. (SATA disks show no PCIe link.)
- **System Health flags downtrained NVMe links.** A new check warns when an NVMe drive negotiates
  fewer PCIe lanes than it supports (almost always motherboard **lane-sharing** — a GPU/second
  card or another M.2 stealing lanes) and notes speed-only reductions as info (a slower slot or
  idle ASPM). The GPU is deliberately excluded — NVIDIA drops its PCIe gen/width at idle, so a
  snapshot would false-alarm.

## [0.37.1] - 2026-05-22
### Fixed
- **`rigdoctor update` now uses the right method for how RigDoctor was installed.** It detects
  apt (`.deb`), pip (venv/`.run`), or source installs (`updates.install_kind()`); only pip
  installs self-update in place. An apt install no longer fails with "No module named pip" —
  it (and the GUI Update button) shows `sudo apt update && sudo apt install --only-upgrade
  rigdoctor`; a source checkout points to `git pull`.

## [0.37.0] - 2026-05-22
### Added
- **Version footer** — a footer across the bottom of the window shows `RigDoctor v<version>` in
  the bottom-right (moved out of the sidebar).
### Fixed
- **Pages scroll when content doesn't fit, and the window is no longer pinned to the tallest
  page's height.** Long pages (Settings, Tuning, …) get a scrollbar when too tall — so controls
  like Uninstall are always reachable — and the window can now be resized smaller than the screen
  (min height dropped from "taller than the screen" to ~600px). Pages that manage their own
  scroll/fill (Dashboard, System Health, Inventory, Share) are unchanged.

## [0.36.1] - 2026-05-22
### Fixed
- `rigdoctor gui` printed the wrong fix when PySide6 is missing — it suggested the non-existent
  `python3-pyside6` package. Now it names the real split modules
  (`python3-pyside6.qt{widgets,gui,websockets,svg}` + `python3-pyte`).

## [0.36.0] - 2026-05-22
### Fixed
- **`.deb` now installs all dependencies automatically — no manual tool install.** The previous
  `Recommends: python3-pyside6` named a package that doesn't exist on Debian/Ubuntu (PySide6 is
  split per module), so apt silently skipped it and the GUI wouldn't start. Now it Recommends the
  actual modules the GUI imports — `python3-pyside6.qt{widgets,gui,websockets,svg}` + `python3-pyte`.
### Changed
- **`apt install rigdoctor` sets up the whole toolset.** The `.deb` also Recommends the optional
  diagnostic/gaming tools (smartmontools, lm-sensors, dmidecode, pciutils, libnotify-bin,
  libsecret-tools, gamemode, mangohud) so they install by default — users never hand-install
  tools. `cpupower` is a Suggests (kernel-tied); `--no-install-recommends` still gives CLI-only.

## [0.35.0] - 2026-05-22
### Added
- **`.deb` package (M9 / D8)** — `packaging/make_deb.py` builds a `rigdoctor_<version>_all.deb`
  (pure-Python, `Architecture: all`) via `dpkg-deb`: `Depends: python3`, with the GUI deps
  (`python3-pyside6`, `python3-pyte`) as **Recommends** so `sudo apt install ./rigdoctor_*.deb`
  gives the full app and `--no-install-recommends` gives CLI-only. Installs the package, both
  launchers, the desktop entry, and the icon. CI (`release.yml`) builds it as a **release asset**
  every release, and optionally publishes it to the Gitea **apt registry** (set a `REGISTRY_TOKEN`
  secret) for `sudo apt install rigdoctor`. **M9 is now complete.**

## [0.34.0] - 2026-05-22
### Added
- **Event-based alerts (M8).** Beyond temperature + GPU-lost, RigDoctor now notifies on
  **critical kernel events** — Xid (GPU error), out-of-memory kills, CPU machine-checks, PCIe
  AER errors, and disk I/O errors — scanned from the kernel log every ~30s while monitoring and
  fired one-shot (cooldown-gated, so no spam). A proactive warning the moment something goes
  wrong, not just on a temperature threshold. Included whenever desktop notifications are on.

## [0.33.0] - 2026-05-22
### Added
- **AI explanations stream live.** "Explain with AI" now fills token-by-token as the model
  generates (Ollama NDJSON + Claude SSE, both via stdlib `urllib`) instead of a multi-second
  freeze, then re-renders the finished answer as Markdown. `core/ai.explain_stream()`.

## [0.32.0] - 2026-05-22
### Added
- **More for diagnostics & reports:**
  - **`nvidia-smi -q` snapshot** — driver, throttle/clock-event reasons, clocks, power, temps,
    PCIe link, ECC + retired pages (point-in-time at diagnostic time).
  - **Display-server log** — auto-detected: `Xorg.0.log` on X11, or the compositor's user-journal
    slice (gnome-shell/kwin/sway/gamescope) on Wayland.
  - **Full system inventory** (M5 hardware/OS) is now included in each stored diagnostic and the
    **Report** bundle — invaluable for larger/shared debugging.
  These join the kernel log + coredump records in `syslogs.txt`/`inventory.*`, are saved per
  diagnostic, included in the Report zip, and (logs) fed to the AI on "Explain".

## [0.31.0] - 2026-05-22
### Added
- **Diagnostics now collect session-scoped system logs** (`core/syslogs.py`): a kernel-log
  slice (`journalctl -k` — Xid, OOM-killer, MCE, PCIe AER, thermal, hung tasks) and
  **crashed-process records** (`coredumpctl` — which executable, signal, and when). They're saved
  to the diagnostic directory (`syslogs.txt`), included in the **Report** bundle, and fed to the
  AI on "Explain" alongside the game logs. Best-effort — degrades quietly if the tools are
  missing or access is denied; scoped to the session window so it doesn't drag in old noise.

## [0.30.0] - 2026-05-22
### Added
- **Logging & report bundles (M15, D25)** — opt-in via one **Settings → Logging** toggle
  (default off). When on: the app logs to a rotating `app.log`, and **each diagnostic is stored
  in its own folder** (`~/.local/share/rigdoctor/diagnostics/<id>/`) with the capture log, a
  structured `result.json`, a readable `report.txt`, a session-scoped game-log snapshot, and an
  `ai/` record of every AI interaction — **the exact data sent, which model, and its reply**.
- **Report** — a button on the diagnostic dialog (and `rigdoctor bundle`) zips a diagnostic's
  folder plus `app.log` into `~/.local/share/rigdoctor/reports/<id>.zip` for sharing. Everything
  stays local; the zip only leaves your machine if you share it. Available only when logging is on.

## [0.29.0] - 2026-05-22
### Added
- **AI now resolves Steam app IDs from your library instead of guessing.** When app IDs appear
  in the logs/findings, RigDoctor looks them up in your scanned games (`steam.appid_names()`) and
  injects an "App IDs (resolved from your installed games)" glossary into the prompt — so the
  model names games correctly (e.g. `2694490 = Path of Exile 2`) rather than hallucinating. Only
  IDs it can resolve locally are listed; no network, no model "training" needed.

## [0.28.1] - 2026-05-22
### Fixed
- **AI explanations were misreading stale/benign logs.** Three fixes so the model analyses the
  *actual* session: (1) the prompt now states the **real game name, capture duration, and
  outcome** (clean vs. crash) so the model stops guessing the game from log paths; (2) game logs
  are **scoped to the session window** (Steam-console lines filtered by timestamp; a stale
  per-app Proton log from an earlier game is skipped); (3) the reference KB flags common
  **benign** Steam/Proton lines (`libnvidia-ml.so.1` assertion, routine minidump uploads, "fork
  without exec") so they aren't reported as the cause. The system prompt also forbids
  Windows-only advice (no "run as administrator") and tells the model not to invent a problem
  when the run was clean.

## [0.28.0] - 2026-05-22
### Added
- **AI explanations now include recent game logs.** When you press "Explain with AI" on a
  diagnostic, RigDoctor also gathers recent **Proton** (`~/steam-<appid>.log`) and **Steam**
  console logs (`core/gamelogs.py`, tail-read + size-bounded) and passes them to the model, so
  it can correlate log errors with the sensor findings and pinpoint *when* something went wrong.
### Fixed
- The AI explanation popup now **renders Markdown** (headings, bold, lists) instead of showing
  raw `###`/`**` — `QTextEdit.setMarkdown`, and the model is told to answer in Markdown.

## [0.27.1] - 2026-05-22
### Changed
- AI assistant: selecting **Ollama** now pre-fills the model field with **`qwen2.5:7b`** (a
  strong 7B that fits an 8 GB GPU; our grounding makes a 7B sufficient). It won't overwrite a
  model you've already entered, and you can change it freely.

## [0.27.0] - 2026-05-22
### Added
- **AI assistant (M14, D24)** — optional, **strictly opt-in, never automatic**. Explains your
  diagnostics in plain language only when you press **"Explain with AI"** on the diagnostic
  dialog (or run `rigdoctor ai explain`). You choose a provider explicitly (no default):
  **Ollama** (local, private, no key) or **Claude** (Anthropic; key stored in the keyring, with
  a consent prompt before any data is sent). Configure in **Settings → AI assistant**.
- Answers are **grounded**: RigDoctor passes the actual findings plus matched reference facts
  from a curated knowledge base (`core/ai_knowledge.py` — exact keyword/code match, no
  embeddings, stdlib only), so even a small local model gets the domain facts it needs. Stdlib
  `urllib` only — no new core dependency. Output is advisory (D9).
- CLI: `rigdoctor ai status|test|explain`.

## [0.26.1] - 2026-05-22
### Fixed
- **Setup wizard contrast.** The **radio buttons** (Recording trigger) were unstyled, so the
  selected option was invisible on the dark theme — now styled with a clear accent ring + dot.
  Bundle **checkboxes** got explicit checked/disabled states, and stay selectable even when a
  bundle is already installed (the page no longer looks dead when everything's present).

## [0.26.0] - 2026-05-22
### Added
- **Graphical setup wizard (M9).** A first-run GUI wizard (`gui/setup_wizard.py`) walks through:
  environment summary → pick **dependency bundles** (Diagnostics / Monitoring / Gaming / Updates,
  from the component catalog) → install the missing apt packages → choose the **recording
  trigger** → a readiness summary. It shows automatically on first launch (until done), is
  re-runnable from **Settings → Run setup wizard** or `rigdoctor-gui --setup`, and `install.sh`
  launches it after a fresh install when a desktop session is present.

## [0.25.0] - 2026-05-22
### Changed
- **Share is now terminal-only (D23, amends D16).** The Share page is a single shared-terminal
  experience: the host shares their shell, the guest watches and may type **only if the host
  ticks "Allow the guest to type"** (otherwise read-only). The terminal is larger and either
  side can pop it **full-screen** (Esc to exit).
### Removed
- The read-only **stats view** (live sensors/health/inventory over the relay) and the
  `rigdoctor share serve` HTTP server — the shared terminal replaces them. (`core/share.py`
  removed; the `share` CLI command is gone.)

## [0.24.0] - 2026-05-22
### Added
- **Shared terminal is now in color.** The terminal view renders pyte's per-cell foreground/
  background, bold, and reverse, so the host's real shell keeps its theming — fish, `ls`,
  `git`, prompts, etc. look the same as locally (the session already runs the host's `$SHELL`
  with its config and `TERM=xterm-256color`; only the rendering was monochrome).

## [0.23.0] - 2026-05-22
### Added
- **Crash-logger trigger modes (M9 / D6)** via `systemd --user`, no root: **manual**,
  **always-on** (a background service records continuously), and **game-launch** (auto-records
  while a Steam game runs). Set it from **Settings → Recording trigger** or
  `rigdoctor service mode <manual|always-on|game-launch>`; `rigdoctor service status` shows it.
  `core/service.py` writes/enables the user units.
- **Zero-config game-launch watcher** (`core/watcher.py`, `rigdoctor watch`) — polls Steam's
  RunningAppID and brackets a focused capture around the running game (the D12 fallback for users
  who don't add the `wrap` launch option; the wrapper stays the precise primary path).

## [0.22.0] - 2026-05-22
### Added
- **M6 breadth.** Environment checks now also report **GPU PowerMizer** mode (NVIDIA, X — flags
  Adaptive/Auto and suggests Prefer-Max-Performance), the **Wine** version, and the **Steam
  client** version.
- **Non-Steam launchers.** Lutris (its SQLite library) and Heroic (Epic + GOG JSON stores) are
  detected (`core/launchers.py`) and listed on the Games page and `rigdoctor games`, tagged by
  launcher. You can Run Diagnostic on them too (records while you play; auto-launch stays
  Steam-only).
### Notes
- The zero-config game watcher (D12 fallback) is deferred to the M9 trigger-mode work, where the
  service integration lives.

## [0.21.0] - 2026-05-22
### Added
- **Live monitor TUI (M2).** `rigdoctor monitor` is now a proper **curses** dashboard:
  current / session-min / session-max per sensor, grouped by subsystem, with temperature and
  utilization **color bands** (and GPU-lost flagged red). `q` quits, `r` resets the session
  min/max. Falls back to a plain full-screen redraw on a non-TTY (`--plain` forces it). The
  terminal face of the same live data the GUI dashboard graphs. Completes the Monitoring bundle.

## [0.20.0] - 2026-05-22
### Changed
- **Reorganized navigation** into grouped sidebar sections — **Monitor** (Dashboard) ·
  **Diagnose** (Games, Recordings, System Health, Tuning) · **System** (Inventory) · **App**
  (Settings, Share) — so it's clear where to go.
- **Renames for clarity:** *Health → System Health* (it's the overall 7-day system scan, not
  per-game), *Environment → Tuning* (gaming tunables + fixes), *Logs → Recordings*,
  *Setup → Settings*.
- **Settings** absorbed **Notifications** (alerts) — app configuration (components/deps, alerts,
  account access, uninstall) now lives in one page; Notifications is no longer a separate item.
- **Recordings** is now a hub: pick which captured log to view (always-on capture, last
  diagnostic, or a preserved crash), **Analyze crash** in place, alongside the recorder controls.

## [0.19.0] - 2026-05-22
### Added
- **System-tray applet (M11, D13).** A tray icon whose menu shows live **CPU / GPU temp** and
  **memory used/total**, a **status line** (Normal / Hot / GPU not responding), and is led by a
  **Run Diagnostic** submenu (pick a detected game → the guided session), plus **Open dashboard**,
  **Start/Stop recording**, **Snapshot (copy)**, and **Quit**. It reuses the dashboard's sample
  stream (no extra sampling). With a tray present, **closing the window hides to the tray** (Quit
  exits); `rigdoctor-gui --tray` starts hidden for autostart. Needs a tray host — on GNOME the
  AppIndicator extension; degrades to a no-op if none is available. Completes the Desktop UI bundle.
- **GUI smoke tests**: construct `MainWindow` headless and exercise the tray, so a startup crash
  fails the build (closes the gap that let the 0.18.0 import regression ship).

## [0.18.2] - 2026-05-22
### Fixed
- **GUI wouldn't start** (0.18.0 regression): the recording indicator used a wrong relative
  import (`from .core` → `rigdoctor.gui.core`, which doesn't exist), crashing `MainWindow` on
  launch. Corrected to `from ..core`.

## [0.18.1] - 2026-05-22
### Changed
- Recording badge: dropped the sample count (not useful at a glance) — it now shows just
  **● Recording** + the game, plus a **⚠ GPU-lost** line if one is detected.

## [0.18.0] - 2026-05-22
### Added
- **Global recording indicator.** While a capture is running, the sidebar shows a red
  **● Recording** badge on every page — with the **game** being captured and the live sample
  count (and a GPU-lost flag if seen). It polls the recorder, so it reflects captures started
  any way: manual `record`, a guided diagnostic, or the Steam launch wrapper.

## [0.17.0] - 2026-05-22
### Added
- **Inventory page is back in the GUI** (it was removed in 0.7.2 in favor of the CLI). Sidebar
  **Inventory** → System / CPU / Firmware / Memory / GPU / Storage / Display as cards, with
  **Copy Markdown** and **Save…** for pasting into forum/bug reports, and **Refresh**. Root-only
  details (motherboard/BIOS/RAM modules via dmidecode) fill in after the launch password prompt.
  Backed by the existing M5 `core/inventory.py` — the CLI `rigdoctor inventory` is unchanged.

## [0.16.0] - 2026-05-22
### Added
- **Automatic crash-capture via a Steam launch wrapper (M6/D12).** Set `rigdoctor wrap
  %command%` as a game's Steam launch option (or in Lutris/Heroic's wrapper field) and RigDoctor
  starts a focused, game-tagged capture when the game launches and stops it cleanly on exit — no
  manual Run Diagnostic / Finish. A hard freeze leaves the capture unterminated, so it's flagged
  as a crash next launch. The wrapper resolves the game name from Steam's `SteamAppId`, doesn't
  disturb an existing capture, and returns the game's exit code. (`core/wrap.py`, `rigdoctor wrap`.)
- GUI **Auto-capture…** helper on the Games page: shows the exact launch-option line (absolute
  path, copy button) and how to set it in Steam.
- Auto-capture preserves an unanalyzed crash (`diagnostic-crash.jsonl`) before starting a new
  capture, so relaunching the game can't wipe a crash report you haven't seen yet.
### Fixed
- `docs/MODULES.md` status column was stale — M1, M3, M4, M5, M8, M10, and M13 are done and now
  marked ✅ (only M2 and M11 remain not-started; M6/M9/M12 in progress).

## [0.15.0] - 2026-05-22
### Added
- **Hard-crash detection & recovery for the guided diagnostic.** If a focused capture ends
  without a clean stop (the recorder never wrote `session-stop` and isn't running), RigDoctor
  treats it as a likely hard freeze. On launch the **Games** page shows a warning banner —
  *"Your last diagnostic for <game> ended unexpectedly…"* — with **Analyze crash** / **Dismiss**.
- **Deeper crash analysis.** *Analyze crash* combines the captured window (final readings before
  the freeze + any GPU-lost event) with a focused scan of the **previous (crashed) boot's kernel
  log** (`journalctl -k -b -1`: Xid/panic/OOM/MCE/AER/thermal) plus SMART/driver/persistence/
  live-temp checks — the full "what happened" picture. `core/diagnostic.py` gains
  `pending_crash()` / `analyze_crash()`; `health.check_previous_boot()` +
  `run_health_checks(include_journal=False)` back it.

## [0.14.0] - 2026-05-22
### Changed
- **Dashboard headline tiles are now history trend graphs** instead of single-value gauges —
  GPU temp, GPU load, CPU temp, and memory each plot their recent history (with the current
  value, window min/max, and a dashed warning-threshold line), so you can see changes over time
  rather than only the instantaneous reading. New `HistoryGraph` widget (QPainter, no new deps).

## [0.13.0] - 2026-05-22
### Added
- **Run Diagnostic now explains itself and can launch the game.** Clicking Run Diagnostic shows
  what to do — *play the game, reproduce the crash, then Finish & analyze* (and that data
  survives a hard freeze + reboot) — and offers **Launch game & start** (asks Steam to run it by
  appid) or **Start without launching**. The recording banner now spells out the next step
  instead of just showing a sample count.
### Fixed
- Button labels containing "&" (e.g. "Finish & analyze") rendered as "Finish _analyze" because
  Qt treated the "&" as a keyboard mnemonic — now escaped so the ampersand shows literally.

## [0.12.0] - 2026-05-22
### Added
- **Guided diagnostic in the GUI.** Each game on the **Games** page now has a **Run Diagnostic**
  button → a focused, game-tagged capture starts and a recording banner appears (live sample
  count, GPU-lost indicator) with **Finish & analyze** / **Discard**. Finishing opens a results
  dialog: the window-scoped capture summary (peak temps/power, events, last samples) plus the
  health findings as cards. The banner persists/restores if you navigate away and back while a
  capture is running. Shares `core/diagnostic.py` with the CLI (one flow, three front-ends).

## [0.11.0] - 2026-05-22
### Added
- **Guided diagnostic session (CLI) — the seed use case, end to end.** `rigdoctor diagnose
  start --game "<name>"` runs a **focused crash-capture tagged with that game** (its own
  diagnostic log, so the report is scoped to just that session), `diagnose status` shows
  progress, and `diagnose finish` stops it and prints a combined report: the **capture
  summary** (peak temps/power, GPU-lost events, last samples — M3) plus the **health findings**
  (Xid/SMART/driver/etc. — M4). The game can be given by `--game` or `--appid` (resolved from
  the Steam scan), and is recorded as a log event so it survives a crash + reboot.
- Shared orchestration lives in `core/diagnostic.py` (one callable for CLI/GUI/tray, per
  ARCHITECTURE §7.1); the recorder/`record run` gained an optional `--game` tag.

## [0.10.2] - 2026-05-22
### Changed
- When an Environment **Apply**/**Install** fails, the status now shows the **real reason**
  (cancelled at the password prompt vs. the system rejecting the change, e.g. a BIOS/kernel-
  locked PCIe ASPM policy) instead of a vague "cancelled, or needs privileges".

## [0.10.1] - 2026-05-22
### Fixed
- **Environment-page contrast.** The combo-box **drop-down list** was rendering light-on-light
  (the popup view is a separate widget the theme didn't cover) — now dark with readable text.
- The **Install / Apply** buttons on findings were hard to read (the accent fill didn't paint
  reliably inside the finding cards, leaving dim dark-on-dark text). They're now an outlined
  style — bright accent text on the dark card, filling accent on hover — readable regardless,
  and given a minimum height so the row can't crush them.

## [0.10.0] - 2026-05-22
### Added
- **Actionable Environment page (M6) — install & apply, not just advice.** Findings that
  recommend a tool or a setting are now one-click:
  - **Install buttons** for GameMode, MangoHud, and cpupower (added to the M9 component catalog,
    so they also appear on the **Setup** page with the existing installer).
  - **Apply controls** for runtime-reversible tunables — a dropdown of the live options + Apply,
    via a single pkexec prompt, no reboot: **CPU governor**, **NVIDIA persistence mode**,
    **PCIe ASPM policy**, **vm.swappiness**, **Transparent HugePages** (`core/fixes.py`). The
    chosen value is validated against the live options before anything runs.
  - This is the consent-gated apply milestone D9 anticipated, scoped to safe settings (**D22**).
    GRUB-based fixes and CPU mitigations stay suggestion-only; `rigdoctor gameenv` still prints
    the exact commands for headless use.
### Changed
- The `Finding` model gained optional `action` (installable component) and `fix` (applyable
  tunable) fields; the shared `finding_card` widget renders the matching control.

## [0.9.0] - 2026-05-22
### Added
- **Gaming environment checks (M6) — the evaluate-and-suggest engine.** A new read-only report
  (D9) that flags system settings which hurt gaming stability/performance and gives the exact fix
  command. Checks: **PCIe ASPM**, **NVIDIA persistence mode**, **CPU governor** (the three that
  map to the seed-case GPU bus-drop / Xid 79), GameMode, MangoHud, `vm.swappiness`, shader disk
  cache, Transparent HugePages, CPU mitigations, and installed Proton versions.
  - **CLI:** `rigdoctor gameenv` (text or `--json`).
  - **GUI:** a new **Environment** page (findings cards, auto-runs on open), reusing the M4
    health-report card style via a shared `finding_card` widget.
### Fixed
- **Notification icon** now uses the RigDoctor icon (matching the app/dock) instead of a generic
  stock icon — resolved from the installed icon theme, the bundled asset, then a stock fallback.

## [0.8.0] - 2026-05-22
### Added
- **Gaming environment checks (M6) — Steam game detection.** RigDoctor now finds your Steam
  libraries (across multiple drives, via `libraryfolders.vdf`) and the games installed in each
  (parsing `appmanifest_*.acf` — stdlib only, no Steam tooling needed). Runtimes, Proton builds,
  and redistributables are filtered out.
  - **Opt-in libraries:** detected libraries are listed with a per-library game count; you check
    the ones to scan. Nothing is scanned until you pick a library.
  - **Background scan on every launch:** the GUI rescans the selected libraries in the background
    when it opens and flags games installed since the last scan with a **NEW** badge plus a count
    on the **Games** sidebar item (cleared when you view the page). Results are cached
    (`~/.local/state/rigdoctor/games.json`) so the list shows instantly.
  - **CLI:** `rigdoctor games` lists detected games; `rigdoctor games libraries
    [--enable PATH | --disable PATH | --all]` lists/selects libraries (headless-complete, D17).
- Config now supports list values (TOML arrays); `steam_libraries` records the selected libraries.

## [0.7.3] - 2026-05-21
### Fixed
- Shared terminal now has **scrollback** — large output (e.g. `ls -la`) can be scrolled up to
  read; it keeps a history buffer and only auto-scrolls to the bottom when you're already there.

## [0.7.2] - 2026-05-21
### Changed
- Removed the GUI **Inventory** tab — use the CLI `rigdoctor inventory` instead. (Inventory is
  still collected for the relay guest view, so a remote helper still sees the host's hardware.)
### Fixed
- Shared terminal caret now sits at the real cursor position (row **and** column) instead of
  the start of the line.

## [0.7.1] - 2026-05-21
### Fixed
- Shared terminal: a guest who joined **after** the host enabled the terminal stayed read-only.
  The host now re-sends the terminal state when a guest joins, so the terminal is available.
- Inventory page no longer jumps back to the top when it refreshes (e.g. when elevated data
  arrives) — scroll position is preserved and unchanged data isn't re-rendered.
- Shared terminal now follows the cursor to the bottom as output arrives (e.g. `ls -la`),
  instead of staying scrolled up.

## [0.7.0] - 2026-05-21
### Added
- **Shared terminal (M12, Tier 3)**: when the host enables it, the session shares a real **PTY**
  shell — the guest gets an interactive terminal (vim, top, tab-completion, Ctrl-C) running on
  the host as the host's user. The host **reads along** live and can type too, e.g. a `sudo`
  password — which stays local and is never sent to the guest. Off by default, host-consented.
  The guest also pulls the host's inventory on join.
### Fixed
- **Input contrast**: all form controls (text fields, spin boxes, combo boxes, terminals) now
  use the dark theme with readable text (Fusion defaulted them to light-on-light).

## [0.6.0] - 2026-05-21
### Added
- **Session sharing over the relay (M12)**: a **Share** tab — *Start shared session* (host)
  hands you a short code and streams a read-only live view; *Enter share code* (guest) joins
  someone else's session and views their sensors/health/inventory. Both connect outbound over
  WebSocket to the relay (`relay_url`, default `wss://rigdoctor.jesseyvanofferen.com`), gated
  by your Gitea access token — no port forwarding. Read-only.

## [0.5.0] - 2026-05-21
### Added
- **Session sharing (M12, Tier 2)**: `rigdoctor share serve` starts a **read-only** live view
  (sensors auto-refresh + health report + inventory) over a local HTTP server, gated by a
  random share token. Bind to localhost for local testing, or to all interfaces behind a
  user-chosen tunnel (Tailscale/cloudflared/SSH) for remote help. No actions, no terminal.
  (Tier 1 export and Tier 3 gated terminal still to come — D16.)

## [0.4.1] - 2026-05-21
### Fixed
- Checkbox contrast: a checked checkbox is now a clear accent-filled box with a checkmark
  (was hard to tell checked from unchecked on the dark theme).

## [0.4.0] - 2026-05-21
### Added
- **Alerts (M8)**: desktop notifications (via `notify-send`) for **overheat** (GPU/CPU past a
  threshold), **GPU-lost** (nvidia-smi timeout), and a **new version available** (fired once
  per version). Edge-triggered with a cooldown so it doesn't spam. Degrades gracefully if
  `notify-send` isn't installed.
- **Notifications page**: configure alerts (enable/disable, GPU/CPU temperature thresholds)
  with a "Send test" button; changes apply live and persist to `config.toml`.
- **App icon**: ships a RigDoctor icon and shows it in the dock/launcher. The GUI
  **self-registers** the icon + `.desktop` on launch (and sets the Wayland app-id), so a
  self-update + relaunch picks it up — no need to re-run the installer.

## [0.3.2] - 2026-05-21
### Changed
- Replaced the per-page "Run with admin" buttons with a **single password prompt at launch**
  (`pkexec`): the GUI collects root-only data (SMART + dmidecode board/BIOS/RAM) once and
  caches it for the session, so Health and Inventory always show the full picture. Falls back
  to non-root if cancelled/unavailable; disable via `elevate_on_launch = false`.

## [0.3.1] - 2026-05-21
### Fixed
- Changelog/release notes now **render Markdown** instead of showing raw `#`/`**` markup —
  the in-app changelog uses `QTextEdit.setMarkdown()` and the update prompt renders notes as
  rich text (closes #1).

## [0.3.0] - 2026-05-21
### Added
- **System inventory (M5)**: CPU, GPU (model/driver/VBIOS/VRAM/PCIe), motherboard/BIOS, RAM
  (total + modules), storage, kernel, and display server. CLI `rigdoctor inventory`
  (`--json` / `--markdown` / `--output`) and a GUI **Inventory** tab with Copy-as-Markdown,
  Save, and "Run with admin" (for `dmidecode` board/BIOS/RAM details). Fills the last GUI tab.

## [0.2.0] - 2026-05-21
### Added
- **"Check for updates" button** in the sidebar — force an immediate version check instead of
  waiting for the 30-minute poll.

## [0.1.1] - 2026-05-21
### Fixed
- Dialogs (the update prompt and changelog) were light-on-light and unreadable — they now use
  the dark theme with readable text.

## [0.1.0] - 2026-05-21
_First milestone release — a complete, installable, self-updating RigDoctor: live monitoring,
crash capture + health report, desktop GUI, user-local install/uninstall, and updates._
### Added
- **In-app uninstaller**: "Uninstall RigDoctor" button on the Setup page (and
  `rigdoctor uninstall [--purge]`) — removes the venv, launchers, and desktop entry, with an
  option to also wipe settings/token/logs. Runs detached so it can delete its own venv.
- **In-app changelog**: a "Changelog" link in the sidebar opens the release history (tags +
  notes) fetched from the update server.

## [0.0.10] - 2026-05-21
### Added
- **"Restart now" button** after a successful in-app update — relaunches RigDoctor for you
  instead of asking you to restart manually.
- **Real release notes**: CI now sets each Gitea release's body from the matching CHANGELOG
  section (instead of "Automated release for…"), and the updater shows **"What's new"** — a
  notes dialog before applying (GUI) and in `rigdoctor update` (CLI).
### Changed
- Setup page / `rigdoctor install`: dropped internal module references (M4, M5, …) from the
  component descriptions — end users don't need them.
- Adopting **Conventional Commits** + **git-cliff** (`cliff.toml`, `packaging/changelog.sh`)
  to generate CHANGELOG entries from commit history going forward (D20).
### Fixed
- The self-extracting **`.run` installer** is now built **without makeself** (a pure-Python
  self-extractor, `packaging/make_run.py`), so it reliably attaches to every release — it was
  silently skipped before because the CI runner couldn't install makeself.

## [0.0.8] - 2026-05-21
### Added
- **Periodic update checks**: the GUI now re-checks for new releases while running (every
  `update_check_minutes`, default 30; 0 disables), so a newly published version is detected
  without restarting. After applying an update, re-checks stop until restart.
- **"Run with admin" on the Health page**: runs all checks (including root-only SMART) via
  `pkexec rigdoctor report --json`, so the full report — not just "SMART needs root" — is
  available from the UI.

## [0.0.7] - 2026-05-21
### Added
- **User-local installer** `install.sh` (no root): creates a private venv, links
  `rigdoctor`/`rigdoctor-gui` into `~/.local/bin`, and adds a desktop entry. Re-run to
  upgrade; `--uninstall` to remove.
- **Self-extracting `.run` installer** via `packaging/make-run.sh` (makeself) — one
  download-and-run executable bundling the wheel + `install.sh`; built and attached to each
  release by CI.
- **Self-update apply (M13)**: `rigdoctor update` now installs the newer version via
  authenticated pip (`rigdoctor[gui] @ git+https://oauth2:<token>@…@<tag>`); the GUI sidebar
  "Update to v…" button applies it and prompts to restart. Token is scrubbed from output.

## [0.0.6] - 2026-05-21
### Added
- **Token-gated updates (M13)**: store a Gitea Personal Access Token, **encrypted in the OS
  keyring** (Secret Service / GNOME Keyring via `secret-tool`) with a 0600-file fallback.
  `rigdoctor login` / `logout` / `update [--check]`; GUI **Setup → Update access** panel
  (token field, "Get a token", backend status) and sidebar states (connect / up-to-date /
  "Update to v…" / access denied). Updates are gated to accounts on the Gitea server (D18).
- `libsecret-tools` added to the installer catalog (enables encrypted token storage).
### Changed
- D18 update mechanism revised from anonymous public HTTP to **authenticated HTTP (token)** —
  the Gitea instance requires sign-in for all anonymous access.

## [0.0.5] - 2026-05-21
### Added
- **M9 installer (first cut)**: detects distro / package manager / GPU; a catalog of optional
  components (smartmontools, lm-sensors, dmidecode, pciutils, libnotify) with what each
  enables; `rigdoctor install [--check] [-y]` installs missing apt packages via pkexec/sudo
  with consent; GUI **Setup** tab with one-click install. Fixes the "smartmontools missing"
  gap in the health report.
- **Update check (M13, check half)**: on GUI launch the sidebar checks the Gitea releases API
  and shows "up-to-date", an "Update to v…" button if a newer release exists, or "update check
  unavailable" if the API can't be reached anonymously.

## [0.0.4] - 2026-05-21
### Added
- **M4 health report**: scans kernel logs (NVIDIA Xid incl. 79 "fell off the bus", kernel
  panic, OOM, MCE, PCIe AER, thermal, amdgpu reset), SMART health, NVIDIA driver/library
  mismatch, journald persistence, and live temps → prioritized plain-language findings with
  suggested fixes (read-only, D9).
- CLI `rigdoctor report` (text + `--json`).
- GUI **Health** tab: runs checks in the background; findings shown as severity-colored cards.
- Tests for the journal scanner.

## [0.0.3] - 2026-05-21
### Added
- Show the app version (`v<version>`) in the GUI sidebar.

## [0.0.2] - 2026-05-21
### Added
- **M3 crash-capture logger**: crash-safe JSONL (`fsync` per sample), size-based rotation,
  GPU-lost/recovered event markers, atomic status file; `rigdoctor record run|start|stop|
  status|report` (foreground `run` is the systemd-ready entrypoint).
- **GUI Recording/Logs page** (M10): start/stop/interval controls, live status, and the
  post-crash report — driving the same recorder via shared `core.reccontrol`.
- Shared render helpers (`format_raw`, `format_headline`, `render_summary`) used by CLI + GUI.
- Tests for the crash log (writer, rotation, reader, summary, recorder).
- **Gitea Actions release workflow** (`.gitea/workflows/release.yml`): on push to `main`,
  builds wheel + sdist and publishes a Gitea release `v<version>` with the artifacts.
### Changed
- **GUI-first** emphasis (D17): docs reframed; the CLI keeps full parity for headless/SSH.
- CPU core temperatures ordered (package, then core 0, 4, 8, …) at the source — fixes the
  CLI ordering too.
- Distribution revised (D8): **user-local self-updating install** is primary, `.deb` optional.
### Planned (docs only)
- M12 session sharing / remote assist (D16); M13 no-root auto-update from the public repo
  (D18); versioning/changelog convention (D19).

## [0.0.1] - 2026-05-21
### Added
- Initial release: planning docs and decisions (D1–D15); **M1 sensor core** (NVIDIA GPU via
  nvidia-smi, CPU via hwmon, memory + DDR5 SPD temps, NVMe); CLI (`snapshot`, `monitor`,
  `sources`); and the **M10 desktop GUI** — dark dashboard with circular gauges and
  collapsible, temperature-colored cards.