test: widen timing margins on two Windows-flaky CI tests

[gnanam1990]

Summary

Investigated two windows-latest smoke-test failures from today's merges. Both are pre-existing test-timing issues, unrelated to the content of the PRs whose merge-to-main runs they failed on (#349, #356 — neither touches internal/specialist or internal/agenteval).

• TestOutputToolBlocksUntilTaskCompletes: a 100ms poll timeout against a 5ms background write left only 20x margin — not enough headroom against Windows' coarser default timer granularity and heavier CI scheduling contention than Linux/macOS. Widened the timeout to 2000ms. This is a ceiling, not a target: PollInterval keeps the test fast in the success case (it still returns as soon as output lands), it just tolerates more jitter before failing.
• TestHarnessTimeoutCancelsBlockedAgent: confirmed there's no actual data race first (harness.Agent.Run is a direct synchronous call on the test's own goroutine, not a spawned goroutine racing the assertions). The failure was workspace materialization alone consuming the full 10s budget on a contended Windows runner, so the agent was never reached before the deadline fired. Widened to 30s. Unlike the other fix, this test's own design blocks for the full timeout to observe cancellation, so this directly costs ~20s more wall-clock time — inherent to what it's testing, not avoidable without restructuring the harness contract.

Test plan

• go build ./...
• go test ./internal/specialist/ -count=1 -run TestOutputToolBlocksUntilTaskCompletes -v (0.01s)
• go test ./internal/agenteval/ -count=1 -run TestHarnessTimeoutCancelsBlockedAgent -v (30.03s, as expected)
• go test ./internal/specialist/ ./internal/agenteval/ -count=1 (full packages)
• go test ./... -race -count=1 (full repo)

Summary by CodeRabbit

• Tests
  • Adjusted automated test timing to better tolerate slower or contended CI environments.
  • Improved timeout handling in long-running test scenarios to reduce flaky failures and keep test runs more reliable.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 9890559b-2ebd-4816-b9de-c58391d8984e

📥 Commits

Reviewing files that changed from the base of the PR and between c0254e3 and ade89a9.

📒 Files selected for processing (2)

• internal/agenteval/benchmark_test.go
• internal/specialist/output_tool_test.go

---

Walkthrough

Two test files had their timeout values increased to reduce flakiness under CI contention: a benchmark harness timeout raised from 10s to 30s, and an output tool test timeout raised from 100 to 2000, both with added explanatory comments.

Changes

Test Timeout Adjustments

Layer / File(s)	Summary
Benchmark harness timeout internal/agenteval/benchmark_test.go	TestHarnessTimeoutCancelsBlockedAgent's BenchmarkInput.Timeout increased from 10s to 30s, with comment explaining insufficient headroom on contended Windows CI runners.
Output tool blocking timeout internal/specialist/output_tool_test.go	TestOutputToolBlocksUntilTaskCompletes's timeout value increased from 100 to 2000, with comments clarifying the timeout acts as a ceiling and returns promptly once output is written.

Estimated code review effort: 1 (Trivial) | ~3 minutes

Possibly related PRs

• Gitlawb/zero#115: Targets the same output-tool blocking/polling behavior whose test timeout is adjusted here.
• Gitlawb/zero#180: Relates to the same benchmark harness timeout/cancellation behavior adjusted here.

🚥 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%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: widening timeouts on two Windows-flaky CI tests.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

---

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

[github-actions]

Zero automated PR review

Verdict: No blockers found

Blockers

• None found.

Validation

• [pass] Diff hygiene: git diff --check
• [pass] Tests: go test ./...
• [pass] Build: go run ./cmd/zero-release build
• [pass] Smoke build: go run ./cmd/zero-release smoke

Scope

Head: ade89a97dd37
Changed files (2): internal/agenteval/benchmark_test.go, internal/specialist/output_tool_test.go

This deterministic review checks validation status and basic diff hygiene. A human reviewer still owns product judgment and design quality.