Repository navigation
fix(ci): restore automated LLM catalog updates - #10735
Merged
Merged
Conversation
Light2Dark
added
enhancement
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Contributor
|
All contributors have signed the CLA ✍️ ✅ |
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
All reported issues were addressed
Reply with feedback, questions, or to request a fix.
Re-trigger cubic
Light2Dark
marked this pull request as ready for review
September 3, 2026 08:29
Light2Dark
enabled auto-merge (squash)
September 3, 2026 08:30
Contributor
There was a problem hiding this comment.
🟡 Changes recommended
The updated catalog includes duplicated Bedrock model entries and the test suite needs small additions to lock in the new denylist + deprecated-filter behavior.
Once you've addressed the issues Copilot identified, you can request another Copilot review.
Pull request overview
Restores the sync-llm-info GitHub Actions workflow’s ability to open/update PRs using the workflow-scoped token, and carries forward a curated LLM catalog sync by improving upstream merge behavior (denylisting + skipping deprecated models) and updating the generated models.yml catalog plus the description-curation skill contract.
Changes:
- Switch PR lookup/creation steps in the weekly workflow back to
GITHUB_TOKENsogh pr list/createhaspull-requests: write. - Extend models.dev ingestion + merge pipeline with a
statusfield and filter out upstreamdeprecatedmodels; expand per-provider denylist. - Update the curated
packages/llm-info/data/models.ymlentries and broaden the description-curation skill to allow narrow factual refreshes.
File summaries
| File | Description |
|---|---|
| packages/llm-info/src/sources/models-dev.ts | Accepts an optional status field from models.dev payloads. |
| packages/llm-info/src/sources/merge.ts | Adds new provider denylist entries and filters out deprecated upstream models during sync. |
| packages/llm-info/src/tests/sync-models.test.ts | Updates denylist assertions; needs additional coverage for new denylist/provider + deprecated filtering. |
| packages/llm-info/skills/SKILL.md | Expands the skill contract to refresh obsolete descriptions, not only fill empty ones. |
| packages/llm-info/data/models.yml | Applies the curated catalog updates (new models, removals, description edits); currently includes duplicated bedrock entries. |
| .github/workflows/sync-llm-info.yml | Uses GITHUB_TOKEN for PR operations to restore automated PR creation. |
Review details
- Files reviewed: 6/6 changed files
- Comments generated: 3
- Review effort level: Lite
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Comment on lines
483
to
487
| - name: Claude Fable 5 (Global) | ||
| model: global.anthropic.claude-fable-5 | ||
| description: "Anthropic's most capable widely released Mythos-class model, built for the most demanding reasoning and long-horizon agentic work" | ||
| description: "Anthropic's Mythos-class model built for demanding reasoning, coding, and long-horizon agentic work" | ||
| roles: [chat, edit] | ||
| capabilities: [thinking, tool_calling] |
Comment on lines
+580
to
584
| expect(MODEL_DENYLIST["opencode-go"]).toEqual( | ||
| new Set(["glm-5.1", "hy3", "minimax-m2.7"]), | ||
| ); | ||
| }); | ||
|
|
Comment on lines
565
to
570
| it("seeds the denylist with unsupported and redundant models", () => { | ||
| // Changing this set is a deliberate product decision — when a model is | ||
| // added or removed here, leave a comment next to it explaining why. | ||
| expect(MODEL_DENYLIST["openai"]?.has("gpt-5.3-codex-spark")).toBe(true); | ||
| expect(MODEL_DENYLIST["openai"]?.has("gpt-realtime-2.1")).toBe(true); | ||
| expect(MODEL_DENYLIST["bedrock"]).toEqual( |
Contributor
Coverage Report for ./frontend
File CoverageNo changed files found. |
kirangadhave
approved these changes
Sep 3, 2026
williambdean
pushed a commit
to williambdean/marimo
that referenced
this pull request
Sep 6, 2026
**This pull request was authored by a coding agent.** ## 📝 Summary The weekly LLM catalog workflow could generate and push updates, but it could no longer open the review PR because a prior catalog update accidentally switched PR operations back to a release-app token without pull-request write permission. This restores the workflow-scoped GitHub token for PR lookup and creation. It also carries forward the stranded catalog update, adds sourced descriptions for all 14 retained new models, and curates provider-specific aliases and older models. Deprecated OpenCode Go and OpenAI entries and an unavailable Azure alias are removed, and future syncs now filter upstream deprecations while denylisting intentionally omitted active models. The description-curation skill also permits narrow factual updates when new models make existing prose obsolete. ## 📋 Pre-Review Checklist - [ ] For large changes, or changes that affect the public API: this change was discussed or approved through an issue, on [Discord](https://marimo.io/discord?ref=pr), or the community [discussions](https://github.com/marimo-team/marimo/discussions). - [ ] Any AI generated code has been reviewed line-by-line by the human PR author, who stands by it. - [x] No visual changes are included. ## ✅ Merge Checklist - [x] I have read the [contributor guidelines](https://github.com/marimo-team/marimo/blob/main/CONTRIBUTING.md). - [x] Documentation has been updated where applicable. - [ ] No new tests were needed; the existing llm-info suite covers schema and generated-output integrity. > Written by gpt-5.6-sol on Codex
This branch was successfully deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request was authored by a coding agent.
📝 Summary
The weekly LLM catalog workflow could generate and push updates, but it could no longer open the review PR because a prior catalog update accidentally switched PR operations back to a release-app token without pull-request write permission.
This restores the workflow-scoped GitHub token for PR lookup and creation. It also carries forward the stranded catalog update, adds sourced descriptions for all 14 retained new models, and curates provider-specific aliases and older models. Deprecated OpenCode Go and OpenAI entries and an unavailable Azure alias are removed, and future syncs now filter upstream deprecations while denylisting intentionally omitted active models. The description-curation skill also permits narrow factual updates when new models make existing prose obsolete.
📋 Pre-Review Checklist
✅ Merge Checklist