Context
-
Project type: !find . -maxdepth 1 ( -name 'pyproject.toml' -o -name 'package.json' -o -name 'Cargo.toml' -o -name 'go.mod' )
-
Test directories: !find . -maxdepth 2 -type d ( -path '/tests/unit' -o -path '/test/unit' -o -path '*/tests/unit' )
-
Last test run: !find .pytest_cache/v/cache -maxdepth 1 -name 'lastfailed'
Parameters
-
$1 : Optional path or test pattern
-
--watch : Enable watch mode for continuous feedback
-
--affected : Only run tests for files changed since last commit
Your task
Delegate this task to the test-runner agent.
Use the Agent tool with subagent_type: test-runner to run fast unit tests only. Pass all the context gathered above and specify Tier 1 (unit tests) execution.
The test-runner agent should:
Run unit tests only (target < 30s):
-
Exclude slow, integration, and E2E tests
-
Use appropriate markers/flags for the framework:
-
pytest: -m "not slow and not integration and not e2e"
-
vitest: --exclude="/e2e/" --exclude="/integration/"
-
cargo: --lib
-
go: -short
Apply options:
-
If --watch : Enable continuous test execution
-
If --affected : Only test changed files
Fail fast: Stop on first failure (-x or --bail )
Provide concise output:
Unit Tests: [PASS|FAIL] Passed: X | Failed: Y | Duration: Zs
Failures (if any):
- test_name: Brief error (file:line)
Rerun failed: [command]
Post-action guidance:
-
If all pass: Ready for continued development
-
If failures: Fix before proceeding (fail fast principle)
-
If > 30s: Suggest /test:consult for optimization
Provide the agent with:
-
All context from the section above
-
The parsed parameters
-
Explicit instruction: Tier 1 only, skip slow tests
The agent has expertise in:
-
Tiered test execution
-
Fast feedback loops
-
Test isolation and parallelization