Skip to content

v0.5.0 changelog

  • Added testing-technique taxonomy covering static/policy checks, white-box unit tests, property/fuzz tests, golden transcripts, snapshots, contract tests, fake integration, black-box smoke, runtime E2E, surface E2E, replay, stress/chaos, security/adversarial testing, semantic eval, and release install smoke.
  • Added composition recipes for runtime + UI evidence, TUI approvals, provider adapters, browser agent safety, channel gateways, scheduler recovery, skill/plugin lifecycle, and release confidence.
  • Expanded latest specification, quickstart, gate matrix, project classification, interaction-surface guidance, best practices, source index, navigation, README, and LLM entrypoints around v0.5.0.
  • Reworded generic standard pages to avoid product-specific framing; product-specific content remains only in legacy compatibility pages, version history, or explicit examples.
  • Bumped package, public plan schema, and public examples to 0.5.0.

Draft standard for evidence-driven quality control of Agent projects.