AgentSkillsCN

Run Slicer Tests

运行 Slicer 测试

SKILL.md

run-slicer-tests

Run the Slicer-based test suite for AdaptiveBrush.

Usage

code
/run-slicer-tests [--exit] [suite]

Where:

  • •--exit - Exit Slicer after tests complete (for automated iteration)
  • •suite is one of:
    • •all (default) - Run all tests
    • •algorithm - Run algorithm tests only
    • •ui - Run UI tests only
    • •workflow - Run workflow tests only
    • •reviewer - Run all Reviewer module tests
    • •reviewer_unit - Run Reviewer unit tests (SequenceRecorder, ViewGroupManager, Bookmarks)
    • •reviewer_ui - Run Reviewer UI tests (navigation, bookmarks, playback, visualization, rating, keyboard)
    • •reviewer_integration - Run Reviewer integration test (full workflow)

What This Skill Does

  1. •Reads SLICER_PATH from .env file
  2. •Launches Slicer with the test runner script
  3. •Executes registered test cases
  4. •Captures screenshots and metrics
  5. •Either exits (with --exit) or leaves Slicer open for manual testing

Prerequisites

  1. •Copy .env.example to .env and set SLICER_PATH
  2. •Ensure the extension is installed in Slicer (symlinked to Slicer's module path)

Execution Steps

Step 1: Read Slicer Path from .env

Use the Read tool to read the .env file and extract the SLICER_PATH value.

Step 2: Launch Slicer with Test Runner

Use Bash to run Slicer with the path from .env:

For automated runs (exits when done):

bash
<SLICER_PATH> --python-script scripts/run_tests.py --exit all

For interactive follow-up (stays open):

bash
<SLICER_PATH> --python-script scripts/run_tests.py all

Replace <SLICER_PATH> with the actual path from the .env file.

Step 3: Check Exit Code

When using --exit, the return code indicates:

  • •0 - All tests passed
  • •1 - One or more tests failed

Step 4: Review Output

Test output is saved to:

  • •test_runs/<timestamp>_<suite>/results.json - Test results
  • •test_runs/<timestamp>_<suite>/metrics.json - Performance metrics
  • •test_runs/<timestamp>_<suite>/screenshots/ - Screenshots per test
  • •test_runs/<timestamp>_<suite>/logs/test_run.log - Test execution log
  • •test_runs/<timestamp>_<suite>/logs/slicer_session.log - Slicer application log

Use /review-test-results to analyze the output (errors, warnings, failures, performance).

Output Location

Test runs are saved to test_runs/ in the extension directory. Each run creates a timestamped folder:

code
test_runs/2026-01-24_143025_all/
├── metadata.json
├── results.json
├── metrics.json
├── screenshots/
│   ├── manifest.json
│   ├── workflow_basic/
│   │   ├── 001.png
│   │   ├── 002.png
│   │   └── ...
│   ├── algorithm_watershed/
│   │   ├── 001.png
│   │   └── ...
│   └── ui_options_panel/
│       └── ...
└── logs/
    ├── test_run.log
    └── slicer_session.log

Interactive Testing (without --exit)

After automated tests complete, Slicer stays open with the Adaptive Brush Tester module visible. You can:

  1. •Start recording manual actions
  2. •Create new screenshot groups
  3. •Take screenshots (auto-numbered within groups)
  4. •Add notes
  5. •Mark pass/fail

Actions are logged to manual_actions.jsonl for later review.

Automated Iteration (with --exit)

Use --exit when iterating on code:

  1. •Run tests: /run-slicer-tests --exit
  2. •Review results: Read test_runs/.../results.json
  3. •Fix issues in code
  4. •Run tests again

This is faster than staying open for manual testing.