feat(tests): add integration, performance tests and CI improvements
- Add 13 integration tests (T-300~T-313) covering engine-platform
message flow, multi-agent coordination, session persistence,
rate limiting, dedup, command registry, role manager, platform
reply, agent permission flow, cron store, card rendering, env
merge, and message attachments
- Add 14 performance benchmarks (T-400~T-413) for message latency,
concurrent throughput, session switching, rate limiting,
deduplication, command lookup, card rendering, cron store,
role resolution, and multi-agent coordination
- Fix FakeAgentSession so NewFakeAgentWithSession properly
preserves pre-configured session for first StartSession call
- Add performance-test Makefile target and update test-release
- Update CI to run performance benchmarks on release events
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Claude
db05dee804