AI agents

Fixtures, mocks, and replay testing

• A Repair Pass When a Checksum Mismatches: Find, Fix or Drop
• Mocking a Memory Store for Fast Policy Tests
• Mocking a Model Endpoint Without Losing Fidelity
• Mocking Tool Responses for Fast Agent Tests
• Pinning Model Versions in Tests: Aliases and Snapshots
• Replaying the Checker: Fixtures From Promoted and Dropped Rows
• Running One Spec With the Cache On and Off
• Simulating Refusals in Staging: Fixtures for Every Code
• Tool Testing: How to Simulate Agent Calls in Staging
• Triaging a Failed Replay: Fold, Retrieval or Prompt
• Why Multi-Step Agent Runs Need Their Own Mock Fixtures
• A Contract Test That Fails on Provider Drift, Not on Your Mock
• A Deterministic Harness for Multi-Turn Agent Tests
• A Fake Billing Backend That Enforces Caps
• A Fake Memory Store That Ages: Injecting Staleness and TTL
• A Fake That Ages: Injecting Staleness and TTL Into Test Fixtures
• A Fixture Format for Recording and Replaying LLM Calls
• A Fixture Format for Recording and Replaying Model Calls
• A Replay Set That Labels Each Failure Class
• A Restore Fixture Set: Replaying Approved and Dropped Chunks
• A Reversal Fixture Built From a Real Refund Trace
• Contract Tests That Fail on Provider Drift, Not on Your Mock
• Fixtures That Survive a Model Swap
• Keeping Fixtures in Sync With Live Policies as Codes Change
• Logging a Mock Hit: Proving a Test Did Not Hit the Real API
• Pinning a Fake Store to a Snapshot of Staging Rows
• Pinning a Field-Order Fixture to a Model Snapshot
• Replaying Recorded Traffic Against a New Agent Version
• Running an Instruction Fixture Suite in CI Without Live Calls
• Seeding a Fake Store With Fixtures From Real Logs
• Seeding a Multi-Step Fixture From a Production Trace
• Seeding Local, CI and Staging From One Definition
• Splitting a Test Suite by Failure Class: Fake, Staging and Live
• Versioning Fixtures When the Checker's Decision Rule Changes
• When Two Workers Race on One Key: Reproducing It