New: Live Test Harness
Test infrastructure for running real Claude Code sessions to verify plugin behavior.
Why Live Tests?
Some behaviors can't be tested with mocks:
- Skill loading patterns
- Natural language trigger matching
- Recursive loop detection
- Claude's interpretation of instructions
Usage
# Run all live tests
make test-live
# Run specific test
bash tests/live/test-prd-skill.sh
# Iterative fix-test loop (wait for fix, re-test)
bash tests/live/fix-loop.sh tests/live/test-prd-skill.shTest Framework
source tests/helpers/live-test-harness.sh
live_test "Test Name" "prompt to claude" \
--timeout 120 \
--max-skill-loads 2 \
--expect "pattern that must appear" \
--reject "pattern that must NOT appear"
live_test_summaryNew Files
| File | Purpose |
|---|---|
tests/helpers/live-test-harness.sh
| Reusable test framework |
tests/live/test-prd-skill.sh
| PRD skill loading tests |
tests/live/test-skill-loading.sh
| General skill tests |
tests/live/fix-loop.sh
| Iterative fix-test automation |
Upgrade
/octo:update --updateFull changelog: https://github.com/nyldn/claude-octopus/blob/main/CHANGELOG.md