Repository navigation
ADFA-6319 | Align the Claude backend plugin with the shared prompt config and capability tags - #122
Merged
Merged
Conversation
…elpers Load the system prompt from YAML assets and reuse ai-core's pane styling and secret reveal. Point help to Agent prefs, fall back on render errors, and drop blank paths from ai-core's IDE context. Refs: ADFA-6319
Daniel-ADFA
approved these changes
Oct 6, 2026
Daniel-ADFA
left a comment
Daniel-ADFA
left a comment
Collaborator
There was a problem hiding this comment.
Reviewed at a8d6a47; no correctness issues found.
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
The Claude plugin worked differently from the other AI backends: its system prompt was hard-coded, the chat couldn't show which Claude model was active, and it kept its own copies of the settings-pane styling and the API-key show/hide control. This change makes it work like the other providers, with the same UX and less duplicate code to maintain.
Technical implementation:
assets/prompts/through plugin-api's shared prompt loader. The loader runs at activation and checks the files. If they can't load or render, the error is logged and AI Core's default prompt is used.LlmBackendRegistration, which keeps it registered across AI Core restarts and reports API-key/model changes. The capability tag uses that to show the active Claude model.IdeContextReadernow leaves out blank paths (the project root itself) from the open-files context.Details
Screen_Recording_20261006_114153_Code.on.the.Go.mp4
Ticket
ADFA-6319