recording-a-known-good-version

To record a known-good baseline version, you capture the accepted version, link the supporting evidence, and define triggers that would require a re-record. This ensures full traceability and trust in your standard changes.

Known-good baseline management is the process of storing the accepted version of changes alongside their supporting evidence and re-check triggers. It ensures that any baseline can be rebuilt from an identified commit, PR, or release.

You can rebuild a baseline by identifying the specific commit, PR, or release where the accepted version was stored. The baseline record links the required evidence and documents any gaps to ensure the version can be reconstructed.

Yes, you need to define re-record triggers for standard changes across prompts, models, tools, dependencies, docs, or lessons. These triggers prompt a re-record when the baseline shifts or when lessons mandate updates.

The best way to handle missing evidence during baseline recording is to document the gaps and use escalation paths. This flags risks and missing evidence to ensure the ship decision is fully traceable and ownership is clear.

Yes, baseline management applies to standard changes across prompts, models, tools, dependencies, docs, or lessons. It links controlled items, ripple effects, and verification to the ship decision and records ownership.