test

Run test suite for current repo. Follow .claude/agents/staff-software-engineer/guides/pipeline.md for retry, approval, publish, and .claude/shared/pipeline-pattern.md for inputs (resolve-mark-proceed) and eval (adversarial).

Print header card before running and footer card after suite finishes. Format: .claude/scripts/stage-card.md.

Detect project test command (in order):

1. Check README.md or CONTRIBUTING.md for explicit test instructions.
2. Maven (pom.xml present): ./mvnw test or mvn test
3. Gradle (build.gradle present): ./gradlew test
4. npm (package.json with "test" script): npm test
5. pytest (pyproject.toml or pytest.ini): pytest
6. Otherwise, infer the runner from the repo: package.json scripts, build files, Makefile targets, CI config (.github/workflows/), or project conventions. If it genuinely cannot be inferred, mark NOT FOUND - NEEDS REVIEW: test runner in the report and proceed (report it). Do NOT ask the user.

Before running, write the phase start marker by running this script. Do NOT inline date/printf (command-substitution + redirect always trips the permission prompt):

.claude/scripts/marker.sh start .claude/runtime/outputs/sse/.markers/{feature_id}.test-generate.start

Run test command. Capture stdout and stderr.

Save .claude/runtime/outputs/sse/test/{feature_id}.md with:

• command run
• exit code
• passed, failed counts
• failing test names (if any)
• duration

Document gates (run on saved report):

• Sensor: .claude/agents/staff-software-engineer/sensors/test-structure.md (auto-run by post-write hook)
• Eval: .claude/agents/staff-software-engineer/evals/test-quality.md (threshold 8.0)

Run the evals adversarially: dispatch a fresh evaluator via the Task tool (subagent_type: general-purpose) that did not author this test report. Hand it only the artifact path and the one rubric path; it scores against the rubrics and reports weighted totals plus the low-scoring dimensions. Below threshold (8.0) retries per pipeline.md, regenerating only the flagged dimensions.

Append approval marker only when exit code is 0 and test-quality eval is >= 8.0. Append with the Edit tool, not Bash: post-eval-sse.sh fires on Edit, and a Bash append skips token accounting and the score log. Keep score= in the shape, phase-log.py only parses markers that carry it:

<!-- approved: {YYYY-MM-DD} score={N} -->

Tests fail, return blocker with failing test names and snippet of failure output. Don't retry automatically; let user decide.

Reply with this exact shape:

Tests {passed|failed}.
  command:  {detected-test-command}
  passed:   {N}
  failed:   {M}
  duration: {seconds}s
  sensors:  test-structure ok
  eval:     test-quality {N}/10
  output:   {path/to/test/output.md}
  next:     /sse:pr (if passed) | fix failing tests (if failed)

Failed, append failures: block listing each failing test name with one-line snippet from failure output.