Add MaxResultSizeHint to ToolAnnotations for agent-level result truncation

[myzie]

Summary

• Adds MaxResultSizeHint int to ToolAnnotations; when > 0, the agent truncates oversized text content blocks after tool execution
• Truncated results are written in full to a temp file (OS-managed cleanup); a footer appended to the result tells the model where to find the full output
• Exports TruncateResult(s string, max int) (result, tmpPath string, err error) for toolkit authors to use directly
• Default hints: ReadFile and Fetch → 100,000 chars; Bash → 50,000 chars
• Truncation is skipped for error results, suspend results, background results, and non-text content blocks

Test plan

• TestTruncateResult/short_string_is_unchanged — no-op below limit
• TestTruncateResult/exact_limit_is_unchanged — boundary case
• TestTruncateResult/over_limit_is_truncated_with_footer — footer contains char count and temp file path
• TestTruncateResult/unicode_is_truncated_by_rune_count — multi-byte chars counted correctly
• TestToolAnnotationsMaxResultSizeHint — JSON round-trip preserves the field
• TestToolAnnotationsZeroHintOmitted — zero value omitted from JSON

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Draft detected.

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.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 83847302-5715-4723-8439-1e5c7cf4cbc3

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

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch worktree-feat+large-result-truncation

---

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