fix: prevent read_file contents from being persisted in ui_messages.json

[roomote]

Related GitHub Issue

Closes: #8690

Roo Code Task Context (Optional)

This PR was created with assistance from Roo Code.

Description

This PR addresses the storage bloat issue where read_file tool contents were being unnecessarily persisted in ui_messages.json, causing excessive storage usage and potential UI failures (gray screen).

Key implementation details:

• Added a new sanitizeMessagesForUIStorage() function that strips large file contents from messages before saving
• The sanitization preserves essential metadata (paths, line ranges) while removing the actual file content
• Implemented backward compatibility via purgeFileContentsFromMessages() to clean up existing bloated messages during rehydration
• Used configurable constants for content truncation thresholds for better maintainability

The solution is non-breaking as:

• API conversation history remains untouched (used for actual rehydration)
• UI still displays file paths and metadata
• Existing tasks with bloated messages are automatically cleaned up

Test Procedure

Testing performed:

1. Added comprehensive unit tests in src/core/task-persistence/__tests__/sanitizeMessages.spec.ts covering:
   • Single file read sanitization
   • Batch file read sanitization
   • Non-file messages preservation
   • Edge cases (empty content, missing fields)
2. All tests passing (11 test cases)
3. Ran existing task-persistence tests - no regressions
4. Ran full Task module tests - all passing

To verify the fix:

1. Create a task that uses read_file on large files
2. Check ui_messages.json - file contents should show [content stripped for storage] instead of full content
3. Rehydrate the task - should work normally (api_conversation_history.json has full content)
4. Load an old task with bloated messages - should be automatically cleaned up

Pre-Submission Checklist

• Issue Linked: This PR is linked to an approved GitHub Issue (see "Related GitHub Issue" above).
• Scope: My changes are focused on the linked issue (one major feature/fix per PR).
• Self-Review: I have performed a thorough self-review of my code.
• Testing: New and/or updated tests have been added to cover my changes (if applicable).
• Documentation Impact: I have considered if my changes require documentation updates (see "Documentation Updates" section below).
• Contribution Guidelines: I have read and agree to the Contributor Guidelines.

Screenshots / Videos

Not applicable - this is a backend storage optimization with no UI changes.

Documentation Updates

• No documentation updates are required.

The fix is transparent to users and developers - it's an internal optimization that doesn't change any public APIs or user-facing behavior.

Additional Notes

This fix significantly reduces storage usage for tasks that read large files. In the reported issue, a ui_messages.json file that was 183MB can now be reduced to just a few KB while maintaining full functionality.

Get in Touch

Available via GitHub for any questions about this PR.

---

Important

Adds message sanitization to prevent large file contents from being stored in ui_messages.json, reducing storage bloat and ensuring backward compatibility.

• Behavior:
  • Adds sanitizeMessagesForUIStorage() in sanitizeMessages.ts to strip large file contents from readFile tool messages, preserving metadata.
  • Introduces purgeFileContentsFromMessages() for backward compatibility, cleaning up existing messages during rehydration.
  • Updates readTaskMessages() and saveTaskMessages() in taskMessages.ts to use new sanitization functions.
• Tests:
  • Adds sanitizeMessages.spec.ts with tests for single and batch file reads, non-file messages, and edge cases.
• Constants:
  • Defines CONTENT_TRUNCATION_LENGTH and STRIPPED_CONTENT_MARKER in sanitizeMessages.ts for content handling.

^{This description was created by for d8c2a0f. You can customize this summary. It will automatically update as commits are pushed.}

[roomote]

Review Summary

I've reviewed the pull request and identified the following issues that should be addressed:

Issues to Address

• Remove unused import of ClineSayTool in sanitizeMessages.ts
• Fix inconsistent sanitization behavior between single and batch file reads (single files preserve content ≤100 chars, batch files always strip content)

Overall Assessment

The PR successfully addresses the core issue of preventing read_file contents from bloating ui_messages.json. The implementation is well-tested with comprehensive test coverage. However, there are minor improvements needed around code cleanliness and consistency in the sanitization logic.

[roomote]

This import is unused. The ClineSayTool type is never referenced in this file.

[roomote]

Inconsistent sanitization between single and batch file reads. Single file reads preserve content if ≤100 characters (line 55), but batch file reads always strip all content regardless of size (lines 66-67). This means the same small file would be saved when read individually but stripped when read as part of a batch, leading to inconsistent UI behavior. Consider applying the same length-based threshold to batch files or documenting this intentional difference.

[ellipsis-dev]

Wrap JSON.parse in try/catch to handle corrupt or malformed ui_messages.json files and log the error to prevent crashes.

^{This comment was generated because it violated a code review rule: irule_PTI8rjtnhwrWq6jS.}