run-tests

name	run-tests
description	Run project tests via subagent to keep main context clean
allowed-tools	Agent, Bash(pnpm test*), Bash(git diff *)

Test Runner

Process

1. Determine test scope:
   • If $ARGUMENTS contains --base-commit <sha>: run git diff --name-only <sha> to get changed files (includes committed, staged, and unstaged changes)
   • Otherwise: run git diff --name-only HEAD to detect changed files (if HEAD is unavailable or no changed files detected, run all tests)
   • Localized changes → run only tests covering changed modules/files
   • Cross-cutting changes (shared utils, config, tsconfig, vite.config, vitest.config) or unsure → run all tests
2. Spawn a subagent (Agent tool) to execute tests
3. Subagent runs the following test commands in order:
   • pnpm test (runs vitest run — full suite)
   • For scoped runs: pnpm test <file-path-or-pattern> (e.g. pnpm test src/utils/__tests__/requestHistory.test.ts)
4. Return the subagent's structured summary to the caller

Subagent Instructions

Execute the test commands listed above. Return a structured summary with one of three statuses:

Status: SUCCESS

• All tests passed
• Per-command results: command, pass/fail

Status: TEST_FAILED

• Per-command results: command, pass/fail
• For failures:
  • Failed test names
  • Error messages
  • Relevant code locations (file:line)
  • Stack trace excerpt (first meaningful lines showing root cause)
• Keep the summary concise but include enough detail to fix the issue without re-running

Status: EXECUTION_ERROR

• Command that failed to execute
• Error output
• This status is for infrastructure/environment errors, not test failures