lucy-ng:supervisor

Route lucy-ng invocations to sanitize, dereplicate, or CASE workflows.

3|2|Updated Jan 8, 2026
One-click install
npx skills add https://github.com/steinbeck/lucy-ng --skill lucy-ng-supervisor
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: lucy-ng:supervisor
Source: https://github.com/steinbeck/lucy-ng/tree/main/skill/supervisor
Command: npx skills add https://github.com/steinbeck/lucy-ng --skill lucy-ng-supervisor

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The lucy-ng Supervisor provides a unified orchestration layer for all lucy-ng invocations, routing requests to sanitization, dereplication, or CASE workflows, and ensuring robust monitoring and intervention when necessary.

Core Features & Use Cases

  • Routes invocations to sanitize, dereplicate, or CASE workflows
  • Spawns and supervises CASE agents, writing CASE-PROGRESS.md after every LSD iteration
  • Detects loop patterns (ELIM thrashing, zero-solution loops, solution explosion, constraint churning) and issues advisory constraints
  • Diagnoses root causes from CASE-PROGRESS.md logs and escalates to the user after 10 failed interventions
  • Collaborates with a diagnostic specialist for deeper root-cause analysis when required
  • Use Case: manage end-to-end CASE workflows in AI-assisted structure elucidation

Quick Start

Use the supervisor to route a lucy-ng workflow: sanitize, dereplicate, or CASE. For a full CASE, initialize a CASE session and monitor progress via CASE-PROGRESS.md.

Frequently Asked Questions about lucy-ng:supervisor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I orchestrate CASE workflows for AI-assisted structure elucidation?

You can orchestrate CASE workflows by using a supervisor to route lucy-ng invocations, spawn CASE agents, and monitor progress through a CASE-PROGRESS.md log after every LSD iteration.

How does an AI agent detect and handle loops in automated CASE workflows?

An AI agent detects loops by monitoring logs for ELIM thrashing, zero-solution loops, or constraint churning, then issues advisory constraints and tracks per-pattern intervention counts to break the cycle.

What is the best way to manage escalation rules when CASE workflows repeatedly fail?

The best way to manage escalation is to enforce intervention limits, delegate to a diagnostic specialist for root-cause analysis, and escalate to the user after 10 failed interventions.

Can I route lucy-ng invocations to sanitize or dereplicate without running a full CASE workflow?

Yes, you can route lucy-ng invocations directly to sanitize or dereplicate workflows independently, bypassing the full CASE session when only specific data processing tasks are needed.

Why does my CASE workflow get stuck during structure elucidation?

Your CASE workflow may get stuck due to solution explosion or constraint churning, requiring an advisory constraint or diagnostic specialist intervention to analyze the root cause from progress logs.