Can I run just a portion of the SDA cycle?

Yes, the skill provides specific commands like '/iterate simulate' for testing only, '/iterate deploy' for canary releases, or '/iterate full' for the complete end-to-end cycle.

What is the SDA loop compared to PDCA?

The SDA (Simulate-Deploy-Augment) loop is an AI-native framework that replaces the traditional Plan-Do-Check-Act (PDCA) cycle. It focuses on testing in sandboxes (Simulate), controlled releases (Deploy), and continuous learning from telemetry (Augment) rather than manual, sequential steps.

How does the SDA Controller handle deployment safety?

It uses a canary strategy that starts with 5% traffic and ramps up over 2 hours. It automatically triggers a rollback if the error rate exceeds 1% or if P99 latency rises above 500ms.

What happens during the Augment phase?

In the Augment phase, the skill analyzes telemetry from sources like Sentry and system logs. It identifies successful patterns to replicate and generates hotfix PRs or configuration updates to address observed failures.

SDA Controller

Name: SDA Controller
Author: Mark393295827

byMark393295827

•

Deployment & DevOps

Optimizes software delivery by replacing traditional PDCA cycles with an AI-native Simulate-Deploy-Augment iteration loop.

The SDA Controller skill introduces an AI-native alternative to the traditional Plan-Do-Check-Act (PDCA) cycle, specifically engineered for high-velocity software development. By transitioning from manual planning to high-fidelity simulation, and from static checking to real-time augmentation, this skill enables Claude to manage the entire lifecycle of a feature. It provides a structured framework for testing solutions in sandboxed environments, executing safe canary deployments with automated rollbacks, and learning from production telemetry to self-correct and improve code quality over time.

Key Features

011 GitHub stars

02High-fidelity simulation sandboxing for pre-deployment validation

03Pre-defined performance thresholds for automated rollback triggers

04Real-time telemetry augmentation that learns from error logs and metrics

05Integrated workflow commands for full-cycle iteration and emergency recovery

06Automated canary release management with configurable traffic ramping

Use Cases

01Accelerating bug-fix cycles through automated simulation and telemetry-driven feedback

02Automating safe canary deployments for microservices with built-in risk mitigation

03Implementing self-healing infrastructure that adjusts configurations based on real-time performance

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mark393295827/house-maint-ai sda-controller

For use in Claude.ai and ChatGPT

Key Features

011 GitHub stars

02High-fidelity simulation sandboxing for pre-deployment validation

03Pre-defined performance thresholds for automated rollback triggers

04Real-time telemetry augmentation that learns from error logs and metrics

05Integrated workflow commands for full-cycle iteration and emergency recovery

06Automated canary release management with configurable traffic ramping

Use Cases

01Accelerating bug-fix cycles through automated simulation and telemetry-driven feedback

02Automating safe canary deployments for microservices with built-in risk mitigation

03Implementing self-healing infrastructure that adjusts configurations based on real-time performance

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mark393295827/house-maint-ai sda-controller

For use in Claude.ai and ChatGPT