Repository navigation
fix(llm): Board agent runs never get the offline demo in place of the real model - #277
Merged
Merged
Conversation
… real model On the lab the member's Claude plan hit its session limit; the failover chain fell over to the Offline client, New idea got its canned "Offline demo specification", and three junk tickets were created on a real repo. Reviewer / Testing calls in the same window received canned text too. An open LlmCallScope already marks a Board agent run (Discovery, Coding, Reviewer, Testing, the Autopilot lead, Retrospective); the factory no longer appends the Offline safety net there, so those runs fail with the real reason and New idea says nothing was created. The Workflow demo / standalone path keeps the fallback. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018WAdQ7Rm24xss5QDSvXtJB
7 of 10 tasks
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
FailoverLlmClientfell over to the keylessOfflineclient every minute (99 times). New idea got the canned "Offline demo specification", and three junk tickets were created on DeltaByteLab/VNChessAcademy (chore(ci): bump azure/login from 2 to 3 in the actions-all group #88–chore(deps): Bump MessagePack from 2.5.301 to 3.1.7 #90, now closed). Reviewer and Testing calls in that window also received canned text.LlmCallScopealready marks a Board agent run (Discovery, Coding, Reviewer, Testing, the Autopilot lead, Retrospective).LlmClientFactoryno longer appends the Offline safety net inside one, so those runs fail with the real reason.Test plan
LlmClientFactoryLiveSettingsTests.Create_InsideBoardAgentRun_NeverFallsOverToTheOfflineDemo: a provider that throws a plan-limit error still falls over to Offline outside a scope, and throws (recorded as a failed call) inside one.dotnet test --solution AgentOs.slnx -c Release: 1,558 passed, 0 failed.🤖 Generated with Claude Code
https://claude.ai/code/session_018WAdQ7Rm24xss5QDSvXtJB