fix(task): mark interrupted tool calls as errors in persisted history

[easonLiangWorldedtech]

Related GitHub Issue

Closes #1283

Description

When a task is interrupted while an assistant turn's tool calls (e.g. an unconfirmed attempt_completion) are still unanswered, resumeTaskFromHistory synthesizes tool_result blocks so the persisted API conversation stays valid. Those synthetic blocks did not carry is_error: true, so the persisted api_conversation_history.json ended with a success-looking completion while the task list recorded the task as interrupted — and the parent stayed delegated. Any consumer that infers completion from the history (imports, tooling, future delegation logic) was fooled into thinking the task finished.

This fix sets is_error: true on the two synthetic tool_result sites in Task.resumeTaskFromHistory:

• assistant-last case: tool results synthesized for every tool_use in the final assistant turn,
• missing-results case: results synthesized for tool_use blocks the final user turn did not answer.

This matches the existing error tool_result pattern already used for new_task truncation in the same file, so the task-list status and the persisted conversation now agree.

Test Procedure

Added two regression tests in src/core/task/__tests__/Task.persistence.spec.ts (describe: "resumeTaskFromHistory interrupted tool calls"). Both spy on initiateTaskLoop to stop the flow right before the agentic loop, feed a persisted history that ends mid-tool-call with initialStatus: "interrupted", and assert the synthetic tool_result blocks passed to the loop carry is_error: true with the correct tool_use_id (the second test also asserts a pre-existing partial result is preserved untouched).

cd src
pnpm exec vitest run core/task/__tests__/Task.persistence.spec.ts   # 15/15 passing (13 pre-existing + 2 new)
pnpm exec eslint --prune-suppressions --max-warnings=0 core/task/Task.ts core/task/__tests__/Task.persistence.spec.ts   # clean, no suppression count changes
pnpm check-types   # 11/11 packages passing

No manual testing required: the change only alters metadata on persisted history and is fully covered by unit tests. No UI changes, so no visual snapshot needed.

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 (see "Test Procedure").
• Visual Snapshot (UI changes only): N/A — no UI changes.
• Documentation Impact: No documentation updates are required.
• Contribution Guidelines: I have read and agree to the Contributor Guidelines.

Get in Touch

easonLiangWorldedtech

Summary by CodeRabbit

• Bug Fixes

  • Interrupted tool calls now correctly appear as errors when a task resumes.
  • Existing tool results and interruption details remain unchanged.

• Tests

  • Added coverage for task recovery in multiple conversation-history scenarios.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 80119019-430c-4b15-a611-acf26181ad60

📥 Commits

Reviewing files that changed from the base of the PR and between 871bb98 and 81c7be9.

📒 Files selected for processing (2)

• src/core/task/Task.ts
• src/core/task/__tests__/Task.persistence.spec.ts

Included review availability: Your plan provides up to 4 included reviews per hour; 1 remains after this review.

---

📝 Walkthrough

Walkthrough

Task resumption now marks synthetic tool_result blocks for incomplete tool calls with is_error: true in both recovery paths. Persistence tests cover these results and confirm that existing tool results remain unchanged.

Changes

Task resumption

Layer / File(s)	Summary
Mark synthesized results as errors src/core/task/Task.ts	Both assistant-last and user-last history recovery paths set is_error: true on synthesized interruption results.
Validate interrupted history recovery src/core/task/__tests__/Task.persistence.spec.ts	Tests cover unanswered tool calls, preserve existing results, and update typing for Anthropic content blocks.

Estimated code review effort: 2 (Simple) | ~15 minutes

Merge Risk: ⚪ Minimal · up to 81c7b

This localized change marks synthesized interrupted tool results as errors and includes regression coverage; no actionable merge-blocking risk remains after normal checks and review.

Suggested reviewers: edelauna

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 1 functions across 2 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The changes satisfy issue #1283 by marking synthetic interrupted tool results as errors in both recovery paths and adding regression tests.
Out of Scope Changes check	✅ Passed	The changes are limited to Task history recovery and related persistence tests, with no unrelated code changes.
Title check	✅ Passed	The title clearly and concisely describes the main fix: marking interrupted tool calls as errors in persisted history.
Description check	✅ Passed	The description explains the issue, implementation, tests, checklist, and documentation impact, with sufficient detail for review.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!