Skip to content

Testing Index

Document ID: TEST-INDEX-001
File Path: docs/15-testing/index.md
Version: 1.0.0
Status: Active
Owner: Quality Engineering Team
Last Updated: 2026-06-27


This document is the central navigation and strategy index for testing the Wovyr AI Platform — the layers of testing, what each guarantees, and how they combine into a confident release pipeline.


╱ chaos · security ╲ (resilience & safety)
╱ performance ╲ (scale & latency)
╱ workflow / e2e ╲ (behavior across services)
╱ integration ╲ (service + datastore)
╱_____ unit (fast, many) ______╲ (logic in isolation)

Most coverage is fast unit tests; fewer, higher-value tests sit above. Determinism is a platform requirement, which makes much of this testable (workflow, memory retrieval).


DocumentResponsibility
unit-tests.mdIsolated logic, fast feedback
integration-tests.mdServices + real datastores
workflow-tests.mdWorkflow & agent behavior end to end
performance-tests.mdLatency, throughput, scale
chaos-testing.mdFailure injection & resilience
security-testing.mdAuth, isolation, supply chain

  1. Deterministic first — exploit the platform’s determinism for reliable tests.
  2. Fast feedback — unit tests run in seconds; gate every PR.
  3. Real dependencies where it matters — integration uses real Postgres/Qdrant/NATS.
  4. Test the contracts — API/DSL/plugin contracts have their own tests.
  5. Safety is tested — isolation and authorization have explicit tests.
  6. CI-enforced — coverage and quality gates in the pipeline.

PR → lint + unit → integration → contract → (nightly: perf · chaos · security)
merge gate

CI uses the CLI and runs against ephemeral Compose stacks.


Per-component targets (see success metrics); critical paths (auth, isolation, workflow execution) aim highest.



VersionDateDescription
1.0.02026-06-27Initial Testing Index