Skip to content

feat: add multi-provider provider intelligence#1

Merged
sir-ad merged 1 commit into
mainfrom
ad/assess-job-posting-integration
Mar 9, 2026
Merged

feat: add multi-provider provider intelligence#1
sir-ad merged 1 commit into
mainfrom
ad/assess-job-posting-integration

Conversation

@sir-ad

@sir-ad sir-ad commented Mar 8, 2026

Copy link
Copy Markdown
Owner

No description provided.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 98e63217e7

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".


const provider = config?.provider || settings.llmConfig.provider
return {
config: buildLLMConfig(settings.llmConfig, secrets, config),

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Respect unsaved API key overrides in provider tests

getActiveRuntimePayload always embeds the stored API key into payload.config, even when the caller passes a new apiKeyOverride. In the background handler, buildLLMConfig gives payload.config.apiKey precedence over payload.apiKey, so Test/Refresh actions keep using the old saved key whenever one exists. This makes key-rotation checks unreliable because users can type a new key in settings and still get diagnostics for stale credentials.

Useful? React with 👍 / 👎.

Comment on lines +130 to +133
if (discovered.length > 0) {
models = discovered
source = "live"
}

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Mark fallback model lists as non-live discoveries

This marks any non-empty discoverModels result as source: "live", but OllamaProvider.fetchInstalledModels returns a built-in fallback list when the endpoint is unreachable. As a result, failed Ollama discovery is reported as a live fetch, which misleads users and can bias smart-default recommendations toward Ollama even when only cloud providers are actually reachable.

Useful? React with 👍 / 👎.

@sir-ad sir-ad merged commit 709ec08 into main Mar 9, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant