Make the review token budget (MAX_TOKENS) configurable — large files are silently skipped and it can't be raised even with a large-context model

[zhu344742542]

A config key in ~/.opencodereview/config.json, e.g. "max_tokens": 200000
(and/or a per-model default keyed off the resolved model)

[lizhengfeng101]

Thanks for the suggestion! Making the token budget configurable is a reasonable idea — it would help users with large-context models get more out of their setup.

Would you be willing to open a feature request issue with a bit more detail? For example:

• Proposed config key name and location (global config, repo-level, or both)
• How it should interact with per-model defaults
• Any thoughts on validation/guardrails (e.g., minimum value, warning when set above the model's actual context window)

That would help us scope and prioritize the work. Looking forward to hearing more!