name	zouroboros-prescribe
description	Self-prescription engine for the Zouroboros system. Takes an introspection scorecard, identifies the weakest subsystem, maps it to a known improvement pattern, and generates a seed YAML + autoloop program.md for autonomous improvement. Includes a governor that flags high-risk prescriptions for human approval.
compatibility	Created for Zo Computer
metadata	{"author":"marlandoj.zo.computer","version":"1.0.0","phase":"Zouroboros Phase 2 — Self-Prescription"}

Zouroboros Prescribe

Takes an introspection scorecard and generates actionable improvement artifacts:

A seed YAML (spec-first format) defining the improvement goal, constraints, and acceptance criteria
A program.md (autoloop format) for autonomous metric optimization
A governor report flagging any risks that require human approval

Usage

bun Skills/zouroboros-prescribe/scripts/prescribe.ts [--scorecard <path>] [--live] [--target <metric>] [--output <dir>] [--dry-run]

Flags

Flag	Description
`--scorecard <path>`	Path to scorecard JSON (from `introspect.ts --json`)
`--live`	Run introspect live and use its output (default if no --scorecard)
`--target <metric>`	Override: prescribe for this metric instead of weakest
`--output <dir>`	Output directory for seed + program (default: `Seeds/zouroboros/`)
`--dry-run`	Show what would be prescribed without writing files

Governor Safety Rules

The governor blocks autonomous execution and flags for human review when:

Prescription touches schema migrations or database structure
Prescription modifies >3 files
Prescription changes executor bridges or routing weights beyond ±10%
Seed ambiguity score > 0.20 (same gate as spec-first-interview)
Target metric has no baseline data (N/A status)

Improvement Patterns

Each metric maps to a known remediation playbook (see references/playbooks.md).

Integration

Input: Introspection scorecard (Phase 1)
Output: Seed YAML + program.md → feeds Phase 3 (autoloop evolution)
Memory: Stores prescription as episode tagged [[zouroboros.prescription]]

Zouroboros Prescribe

Takes an introspection scorecard and generates actionable improvement artifacts:

A seed YAML (spec-first format) defining the improvement goal, constraints, and acceptance criteria

A program.md (autoloop format) for autonomous metric optimization

A governor report flagging any risks that require human approval

Usage

bun Skills/zouroboros-prescribe/scripts/prescribe.ts [--scorecard <path>] [--live] [--target <metric>] [--output <dir>] [--dry-run]

Flags

Flag

Description

--scorecard <path>

Path to scorecard JSON (from introspect.ts --json)

--live

Run introspect live and use its output (default if no --scorecard)

--target <metric>

Override: prescribe for this metric instead of weakest

--output <dir>

Output directory for seed + program (default: Seeds/zouroboros/)

--dry-run

Show what would be prescribed without writing files

Governor Safety Rules

The governor blocks autonomous execution and flags for human review when:

Prescription touches schema migrations or database structure

Prescription modifies >3 files

Prescription changes executor bridges or routing weights beyond ±10%

Seed ambiguity score > 0.20 (same gate as spec-first-interview)

Target metric has no baseline data (N/A status)

Improvement Patterns

Each metric maps to a known remediation playbook (see references/playbooks.md).

Integration

Input: Introspection scorecard (Phase 1)

Output: Seed YAML + program.md → feeds Phase 3 (autoloop evolution)

Memory: Stores prescription as episode tagged [[zouroboros.prescription]]

zouroboros-prescribe