Repository navigation
Conversation
The model client implements the Google provider while the guide also advertises an unsupported Anthropic configuration. Correct the manifest guidance and exercise its actual model YAML through the API, store, and model provider client.
erain
force-pushed
the
docs/supported-model-providers-371
branch
from
November 10, 2026 16:50
e5f02e2 to
875c042
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Applying the documented Anthropic Model succeeds, but the workspace planner fails with
unsupported provider. The guide now documents the implemented Google provider and removes the unsupported example. It also explains Model resource scope and secret namespace, corrects the supported-kind count and Task image requirements, and replaces a broken example link.Fixes #371.
Validation on Go 1.27.1, Linux/amd64 (all exited 0):
Prepared with an AI coding assistant.
Regression output, API/provider fixture, and validation limits
A regression test reads the actual Model YAML from
docs/manifests.md, sends it through a real loopback gRPC AX server and memory store, then runs the workspace planner against a local Google-format HTTP fixture. It checks the saved spec, secret reference, request method/path/authentication, workspace goal/system instruction, exact returned plan, request count, and fixture cleanup. An Anthropic negative control verifies API acceptance followed by generation failure with zero provider requests.Reproduction and captured behavior
Before the documentation correction, on base
ac2332829f22360ff97b0ba34d94dd0dd782f17ewith the regression test added:After the correction:
Validation
On Go 1.27.1, linux/amd64, all of these completed with exit status 0:
The full test suite contained no skipped tests. Checked all local links in the edited guide and reviewed the complete diff, including the new test.
The provider fixture runs entirely on loopback with a synthetic secret resolver; this is API/store/planner/client integration evidence, not a live Google service, Kubernetes Secret, paid-model, or Substrate sandbox test. It does not establish external availability of the example model ID. The current Task runner has its own Gemini bootstrap configuration; Model resources do not automatically configure arbitrary agent frameworks. The image digest and API key placeholders require user substitution.