Repository navigation
feat(meta): integrate Muse Code through the shipped Meta provider - #5
Merged
Merged
Conversation
DeepSeek Harness already ships the `meta` pi-ai route and its device-code
flow, so the bundle now integrates them instead of inserting a second
`@deepseek-ai/dsh-llm-pi-ai` instance, which cannot activate because every
instance declares the whole provider catalog.
- declare the `meta` route in the `llm-pi-ai` settings section so the
provider appears under Settings > Models and in the model picker
- add the `/muse-login` command and a Sign in control under Settings > Models,
backed by an exact local `POST/GET /muse-code/login` route
- open the Meta verification page in the default browser and show the code
- read the authorization seam's neutral notice shape (`message`/`url`/`code`)
- ship the Client half as a browser classic script; the ESM `export {}` that
tsc emitted aborted the Web boot
- drop the hand-written route, `MUSE_CODE_SUB_TOKEN` handling, and the
`muse-wonder` ids that DSH's catalog does not serve
BREAKING CHANGE: 0.1.x embeddings and their credentials are replaced; see
CHANGELOG and README for the 0.2.0 upgrade note.
TheStreamCode
force-pushed
the
feat/meta-provider-integration
branch
from
October 26, 2026 09:42
870bf07 to
42e88f2
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Reworks the bundle onto the
metaprovider that DeepSeek Harness already ships, replacing the unsupported second@deepseek-ai/dsh-llm-pi-aiinstance.muse-code-llmrow could not activate in DSH 0.2.1-rc.3 (configurable provider "amazon-bedrock" is already declared), so nothing appeared under Settings > Models or in the picker (/muse-loginwas added in this branch as an unreleased fix).metaroute in thellm-pi-aisettings section, adds/muse-login, and adds a Sign in control under Settings > Models backed by an exact localPOST/GET /muse-code/loginroute (DSH exposes no Remote method for provider OAuth flows).client/client.js); the previouslib/client.jscarriedexport {}from tsc and aborted the Web boot withUnexpected token 'export'.Notes
MUSE_CODE_SUB_TOKENhandling, and themuse-wondermodel ids that DSH's catalog does not serve.AUTH_CONTRACT.mdkeeps the pinned wire values; the text now records that DSH implements the flow.lib/rebuilt fromsrc/.Verification
npm run typecheck,npm run buildnpm test(17 tests: notice shapes, command, HTTP route GET/POST/409/504, client artifact classic-script parse and click path, manifest)markdownlint-cli2on the docs workflow files: 0 issuesmetaregistered as Muse Code (Meta subscription), catalog models listed,llm-pi-ai/metaflow present,muse-logincommand registered, no activation warning