zouroboros-prescribe

To convert an introspection scorecard into remediation artifacts, apply a prescriber workflow that identifies the weakest subsystem, maps it to a remediation playbook, and generates a seed YAML plus an autoloop program for autonomous improvement.

Governor safety gating flags high-risk prescriptions for human approval by enforcing safety rules against schema, file, and routing changes, ensuring that ambiguous or hazardous autonomous improvement cycles are intercepted before execution.

Automating self-healing workflows involves running a prescriber on live introspection or JSON scorecards to identify subsystem regressions, map them to playbooks, and emit an autoloop program to orchestrate automated metric optimization and repair.

Yes, you can use live introspection feeds instead of JSON scorecards. The prescriber accepts both formats to identify the weakest metric, enforce governor safety rules, and output the required seed YAML and program.md files for the autoloop.

A self-healing prescriber outputs a spec-first seed YAML defining goals and constraints, a program.md formatted for the autoloop, and a governor report detailing safety gating for high-risk schema, file, or routing changes.

You should avoid automated seed YAML generation when a prescription triggers governor safety rules for high-risk schema, file, or routing changes, requiring human approval before the autoloop program can safely execute autonomous improvements.