running-tests

Prevent opencode-swarm session crashes by using three-layer session guardrails that restrict unbounded test fan-out when executing the test runner tool. This avoids scope_exceeded errors and session kills during targeted test execution.

Mock module cache poisoning occurs when tests execute without proper isolation. Cross-platform bash and PowerShell per-file test loops provide test isolation, ensuring each test runs independently without contaminating shared module caches.

Classify test regressions by following a failure triage workflow that uses Git worktrees to verify pre-existing failures, parses CI logs, and identifies whether failures existed before your code changes to avoid false bug reports.

Run isolated tests for a single opencode-swarm agent file by applying scoped test runner configuration with cross-platform per-file test loops, ensuring only relevant tests execute without triggering session instability or scope_exceeded errors.

Parse CI logs as part of the failure triage workflow to identify pre-existing test failures before reporting regressions. This structured approach verifies whether failures existed prior to your code changes, preventing false bug reports.

The skill supports both bash and PowerShell for cross-platform test isolation, providing per-file test loops that prevent mock module cache poisoning and ensure reliable test execution across different operating system environments.