feat: add Mistral API parameters

[guillaumegay13]

Summary

• add Mistral API-key parameter rows for 13 commonly used chat models
• use exact Mistral /v1/chat/completions request paths, including nested response_format.type
• cover generation length, sampling, output format, and Mistral safe prompt controls

Docs checked

• https://docs.mistral.ai/api
• https://docs.mistral.ai/openapi.yaml
• https://docs.mistral.ai/models/overview

Notes

• reasoning_effort is intentionally not included in this PR because current Manifest routing still strips it as an OpenAI-only field for non-OpenAI providers. That should be handled in a separate adapter/routing PR before publishing it in MPS.
• response_format.type exposes text and json_object; json_schema needs the companion response_format.json_schema object, which MPS cannot model yet as a primitive parameter.

Validation

• npm run validate
• npm test
• npm run typecheck
• npm run lint
• npm run build
• npm run guard:params -- --base upstream/main
• npx prettier --check .
• git diff --check
• live Mistral /v1/chat/completions smoke from the new YAML paths returned HTTP 200 for every added model

[vercel]

@guillaumegay13 is attempting to deploy a commit to the Manifest Team on Vercel.

A member of the Team first needs to authorize it.