fix: have option for cli providers to use their configured or default model #3683
Conversation
| .arg("json"); | ||
| .arg(&filtered_system); | ||
|
|
||
| // Only pass model parameter if it's in the known models list |
There was a problem hiding this comment.
Consider adding
// This allows users to use "default" or any future CLI models without updates
(and same for gemini_cli.rs)
|
@michaelneale Great solution! I think as is your solution is excellent, main feedback I could give is consider updating We could either:
What do you think? |
|
ideally if someone types it in then we would pass it through - otherwise we dont' pass anything, maybe that is better approach vs having any list/hard coding at all? |
|
I still think having a known models list makes sense to prevent user error that could cause the CLI to fail (ie. invalid/outdated/unsupported model or typos etc). The middle ground would be to only skip passing the model for specific values like "default" or empty string, otherwise let it pass through. That would avoid hardcoding model lists while allowing CLI defaults. Either way, what you have in this PR should be sufficient to solve the regression issue, I think it's a great solution as is. |
* main: blog: streamlining detection development w/ recipes (#3689) fix: have option for cli providers to use their configured or default model (#3683) docs: new blog post and corrections to an old one on goosehints (#3657) Resolve sub recipe path relative to the parent recipe path (#3642) Speed up recipe loading from deeplinks and various fixes (#3662) fix cmd + , not opening settings (#3694) Add warning when JSON env parsing fails. (#3696) chore: refactor session naming into provider (#3678) feat (ui): File picker for scheduling recipes default to recipe dir (#3611) fix: address issue with streamable http interactions via mcp (#3693) Provider scenario tests (#3688) Fix conversations before they hit the LLM (#3660) cli: add detailed instruction for WSL users (#3496) feat: recipe runs will now prompt for missing extension secrets (#3668) fix: pricing integration tests -> trying more runs for cache and retries (#3546)
* main: fix: Ensures final output tool is available when using vector tool search (#3701) chore: adding in some new models token limits (#3685) blog: streamlining detection development w/ recipes (#3689) fix: have option for cli providers to use their configured or default model (#3683) docs: new blog post and corrections to an old one on goosehints (#3657) Resolve sub recipe path relative to the parent recipe path (#3642) Speed up recipe loading from deeplinks and various fixes (#3662) fix cmd + , not opening settings (#3694) Add warning when JSON env parsing fails. (#3696) chore: refactor session naming into provider (#3678) feat (ui): File picker for scheduling recipes default to recipe dir (#3611) fix: address issue with streamable http interactions via mcp (#3693) Provider scenario tests (#3688)
… model (block#3683) Signed-off-by: Adam Tarantino <[email protected]>
fixes: #3682