Can I manually check the health of my gates?

Yes, you can trigger an on-demand health check at any time by asking Claude 'how are the gates doing?'

Which metrics are tracked by this skill?

It monitors time-to-gate, mode selection accuracy, sentinel risk levels, surgeon agreement, gate outcomes, override rates, and LLM costs.

What is the primary purpose of Invariance Health?

It acts as a self-evaluating feedback loop for quality gates, ensuring the system doesn't become too restrictive (causing delivery freezes) or too lenient (allowing bugs into production).

How does the 10-invocation retrospective work?

Every 10 gate runs, the skill queries the evidence store, calculates aggregate metrics, and uses specialized LLM roles to analyze whether thresholds need adjustment.

Invariance Health Metacognition

Name: Invariance Health Metacognition
Author: supportersimulator

bysupportersimulator

0•

분석 및 모니터링

Optimizes quality gate effectiveness by monitoring metrics and auto-calibrating LLM consensus thresholds to prevent development bottlenecks.

Invariance Health serves as the metacognitive layer for the 3-Surgeons consensus system, providing a critical feedback loop that ensures quality gates remain accurately calibrated. It tracks deep performance metrics including surgeon agreement, override rates, and false positives/negatives across gate invocations. By conducting automated retrospectives every ten cycles, the skill utilizes multi-model cross-examination to identify patterns of over-gating or quality erosion, recommending empirical adjustments to risk thresholds and mode selection to balance delivery velocity with production safety.

주요 기능

010 GitHub stars

02Comprehensive metric tracking including latency, cost, and surgeon agreement

03Data-driven threshold adjustment recommendations for risk stratification

04Real-time anomaly detection for frequent gate overrides or time-to-gate spikes

05Automated 10-invocation retrospective analysis and reporting

06Multi-LLM pattern classification for over-blocking or under-blocking detection

사용 사례

01Fine-tuning risk thresholds after identifying escaped bugs or false positives

02Calibrating automated code review gates to reduce developer friction

03Auditing the performance and cost-efficiency of multi-model LLM workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add supportersimulator/3-surgeons invariance-health

For use in Claude.ai and ChatGPT

주요 기능

010 GitHub stars

02Comprehensive metric tracking including latency, cost, and surgeon agreement

03Data-driven threshold adjustment recommendations for risk stratification

04Real-time anomaly detection for frequent gate overrides or time-to-gate spikes

05Automated 10-invocation retrospective analysis and reporting

06Multi-LLM pattern classification for over-blocking or under-blocking detection

사용 사례

01Fine-tuning risk thresholds after identifying escaped bugs or false positives

02Calibrating automated code review gates to reduce developer friction

03Auditing the performance and cost-efficiency of multi-model LLM workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add supportersimulator/3-surgeons invariance-health

For use in Claude.ai and ChatGPT