fix: token counting issues in ChatAgent context management

[Wendong-Fan]

Description

• Fix token_limit property to correctly use user-provided value instead of always comparing with model's limit
• Reset summary state when clearing memory to prevent stale token counts
• Re-fetch context after summarization to use accurate token counts
• Pass token_limit to context summarizer agents for consistent behavior
• Remove reactive token limit error handling in favor of proactive summarization
• Improve warning message in UnifiedModelType.token_limit to show model name

Checklist

Go over all the following points, and put an x in all the boxes that apply.

• I have read the CONTRIBUTION guide (required)
• I have linked this PR to an issue using the Development section on the right sidebar or by adding Fixes #issue-number in the PR description (required)
• I have checked if any dependencies need to be added or updated in pyproject.toml and uv lock
• I have updated the tests accordingly (required for a bug fix or a new feature)
• I have updated the documentation if needed:
• I have added examples if this is a new feature

If you are unsure about any of these, don't hesitate to ask. We are here to help!

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch token_count_fix

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[fengju0213]

Everything looks great, apart from the missing guard condition I mentioned.

I also have a suggestion regarding extensibility. Since we've completely removed the fallback mechanism in this PR, I wonder if we should consider retaining a limited fallback for cases where the context window is exceeded despite our proactive measures. Even if we trigger summarization at 50% context usage, there are edge cases where a single turn could still overwhelm the entire context window.

Previously, we handled this by providing a condensed summary of the tool call that caused the overflow, which gave the agent feedback to avoid repeating the same strategy. Since this logic has been removed, perhaps we could consider re-introducing a strategy (possibly in a future PR) to truncate tool outputs when a context explosion occurs, while still preserving a summary of the cause for the agent. What are your thoughts on this?