evaluateskills-postmortem

A post-mortem audit compares mutated skills against their pre-run baseline to detect narrative drift. It helps maintain the integrity of skill outputs by identifying regressions after evaluations.

You can detect narrative drift by running a post-mortem audit against the pre-run baseline. This process highlights narrative changes that could impact output quality and provides actionable alerts.

Yes, the system conducts parallel audits using background agents for efficiency. You can initiate this automated post-mortem review directly by executing the `/evaluateskills-postmortem` command in Claude Code.

After detecting narrative regressions, the audit provides clear alerts and decision-making support. This allows you to review the outcome and choose whether to restore the pre-run baseline or keep the mutated changes.

The automated post-evaluation review focuses specifically on detecting narrative drift and regressions. It requires a pre-run baseline to function and does not handle general skill development tasks outside of post-mortem auditing.

You need narrative critical analysis after skill evaluations to ensure narrative integrity. It is necessary when mutated skills undergo changes that might introduce regressions and degrade the quality of the outputs.