llama.cpp-cu133(-dll): Add version b9404

[NecroMan]

Updated llama.cpp for CUDA 13 to CUDA 13.3 version.
Previous CUDA 13.1 version is no longer being updated.

Summary by CodeRabbit

• Chores
  • Added Windows x64 CUDA 13.3 package manifests for llama.cpp (DLL and standard binary distributions).
  • Updated packages to the latest release (b9404) and ensured download verification and model persistence paths.
  • Enabled automatic version discovery and autoupdate from upstream releases.
  • Environment path setup included so installed binaries are immediately usable.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 3d985e60-5165-4a29-b0bf-7ecf83ccf0ac

📥 Commits

Reviewing files that changed from the base of the PR and between 56af95a and ea0a0f9.

📒 Files selected for processing (2)

• bucket/llama.cpp-cu133-dll.json
• bucket/llama.cpp-cu133.json

🚧 Files skipped from review as they are similar to previous changes (2)

• bucket/llama.cpp-cu133-dll.json
• bucket/llama.cpp-cu133.json

---

Walkthrough

Adds a new Windows x64 CUDA 13.3 Scoop manifest and updates the DLL manifest: both pinned to version b9404 with corresponding download URLs, SHA-256 hashes, GitHub-based checkver regex, and $version autoupdate templates.

Changes

llama.cpp CUDA 13.3 Package Support

Layer / File(s)	Summary
DLL distribution manifest bucket/llama.cpp-cu133-dll.json	Updated JSON manifest: version b9404, replaced 64-bit DLL download URL and SHA-256 hash; env_add_path, GitHub checkver regex, and $version-templated autoupdate remain.
Full release manifest bucket/llama.cpp-cu133.json	New JSON manifest for Windows x64 CUDA 13.3: version b9404, release ZIP url and hash, bin executable list, persist: models, GitHub checkver regex b[\d]+, and $version autoupdate template.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• ScoopInstaller/Versions#2761: Adds/updates similar llama.cpp CUDA manifests with comparable structure and autoupdate patterns.

Suggested labels

review-needed

Suggested reviewers

• z-Fng

Poem

🐰 I found two manifests, snug and new,
Bumped to b9404 with hashes true,
Paths and bins in tidy rows,
GitHub tags will keep them chose,
Hooray — the rabbit ships them through!

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Description check	❓ Inconclusive	The description mentions the CUDA version upgrade and context about the previous version, but lacks the conventional PR title format and issue references required by the template.	Add issue reference (Closes #XXXX or Relates to #XXXX) and follow conventional format: '<manifest-name[@version]|chore>: '.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'llama.cpp-cu133(-dll): Add version b9404' clearly identifies the manifests being updated and the version bump, which matches the main changes in the changeset.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@bucket/llama.cpp-cu133-dll.json`:
- Around line 13-16: Replace the current checkver block that uses "url" and
"jsonpath" with Scoop's GitHub-mode by adding "checkver.github" and set it to
the repository identifier (e.g., "ggml-org/llama.cpp") — do not include
"/releases" — and keep the existing "regex" if still needed; remove the manual
"url" and "jsonpath" keys so Scoop will use its token-authenticated GitHub API
path. Ensure you update the "checkver" object to use the "github" key instead of
"url"/"jsonpath" while retaining the same "regex" value.

In `@bucket/llama.cpp-cu133.json`:
- Around line 32-35: The checkver block currently uses a generic url/jsonpath
pair; replace it with Scoop's GitHub mode by changing the "checkver" object to
use the "github" key pointing to the repository (e.g. "ggml-org/llama.cpp") and,
if needed, specify the release/tag source or asset selector instead of "url" and
"jsonpath"; update any "regex" to match the repo tag format and remove the
now-unnecessary "url" and "jsonpath" entries so checkver uses GitHub API
auth/rate-limit handling.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: a13e90bd-df52-41a3-96e0-347164fafbeb

📥 Commits

Reviewing files that changed from the base of the PR and between 4cb9abb and c3a3aeb.

📒 Files selected for processing (2)

• bucket/llama.cpp-cu133-dll.json
• bucket/llama.cpp-cu133.json

[NecroMan]

When I use checkver.github, checkver.ps1 returns an invalid build.

[z-Fng]

/verify

[z-Fng]

When I use checkver.github, checkver.ps1 returns an invalid build.

#2865 (comment)

[github-actions]

All changes look good.

Wait for review from human collaborators.

llama.cpp-cu133-dll

• Lint
• Description
• License
• Hashes
• Checkver
• Autoupdate

llama.cpp-cu133

• Lint
• Description
• License
• Hashes
• Checkver
• Autoupdate

Check the full log for details.

[z-Fng]

Thank you for your contribution!