Add tested support for MSI Alpha 17 C7VF / C7VG

[Liuwins]

Summary

Add tested support for the MSI Alpha 17 C7VF / C7VG on EC firmware 17KKIMS1 based on the owner's reports in issues #151, #152, and #153.

Type of change

• Hardware or model support
• Documentation
• Build, tooling, or CI change

Related issue or discussion

• Closes [Model] MSI Alpha 17 C7VF / C7VG (17KKIMS1.115) #151
• Closes [Curve] MSI Alpha 17 C7VF / C7VG (17KKIMS1.115) #152
• Closes [Power] MSI Alpha 17 C7VF / C7VG (17KKIMS1.115) #153

Changes

• Promote 17KKIMS1 from Experimental to Tested.
• Register the confirmed single-byte tachometers at 0xC9 / 0xCB.
• Mark the confirmed fan-curve tables at CPU 0x72 / GPU 0x8A as verified.
• Use the captured 0xD2 / 0xD4 profile recipes. 0xEB stayed 0x00 in all four MSI Center scenarios, so the Alpha profile deliberately does not write an unverified Super Battery register.
• Keep the existing 99 °C-for-5-seconds safety stop. Issue [Power] MSI Alpha 17 C7VF / C7VG (17KKIMS1.115) #153 is recorded as thermal saturation rather than used as a reason to weaken the protection or claim a complete power ranking.
• Regenerate data/models.json and docs/SUPPORTED_MODELS.md, update the model counts, and add a changelog entry.

User-visible changelog

• MSI Alpha 17 C7VF / C7VG is now listed as tested.
• Fan RPM and verified CPU/GPU fan-curve editing are available for 17KKIMS1.
• Super Battery does not write 0xEB on this model because MSI Center left it at 0x00 in every captured scenario.

Contribution authorship and code origin

• Created by the submitter specifically for GhostDeck
• Created with contributions from other people

The register values and hardware observations come from Liuwins' read-only reports in issues #151, #152, and #153. No code was adapted or ported from another project.

AI assistance

• This pull request was created with generative AI assistance
• Codex

AI was used for code generation, code review, debugging, documentation, and research/explanation. The submitter reviewed the resulting diff and remains responsible for the contribution.

Testing

Test environment

• Windows development environment
• .NET SDK 8.0.424
• No live EC writes or hardware tests were performed by this submission process.

Checks performed

• dotnet build GhostDeck.csproj -c Release --no-restore — 0 warnings, 0 errors
• GhostDeck.dll --verify-models data/models.json — 149 models, round-trip byte-identical, matches compiled tables
• python tools/lang-check.py — 701 keys translated in all 15 languages, no duplicate keys
• Alpha model assertions — 7/7 passed, including Tested tier, 0xC9/0xCB RPM, verified 0x72/0x8A curve, and no 0xEB=0x0F write

Maintainer follow-up

data/models.json is intentionally updated together with the compiled tables, but its detached signature cannot be regenerated here because the repository signing private key is not available to the contributor. Before merging, please run pwsh tools/sign-models.ps1 with the repository's original signing key and commit the resulting data/models.json.sig. The public key and trust root are unchanged.

[wygodad]

@Liuwins Thank you for this - a careful, well-built PR, and the first outside contribution to the model database. It arrived in parallel with our own work on your reports: everything it proposes is already on main as of today (commits 14a4ab5 and e860420) - the tested tier, the tachometers at 0xC9/0xCB, the verified curve, and notably the same call you made independently on 0xEB: your capture in #151 showed MSI Center never touches it on this board, so the recipe skips that write. You are in the entry's credit, and the update is live in the signed model database.

Closing this PR as superseded rather than merging: the branch now conflicts with main in all five files, and data/models.json plus its signature have to be regenerated with the repository's signing key on the maintainer's side anyway, exactly as you noted. Thanks again - report sets like your #151-#153 are what moves this database forward.