How does this skill help reduce token costs?

It features an interactive configuration phase where you select only the specific analyses you need, and its efficiency analysis identifies wasteful patterns like redundant operations.

What are the reporting styles available?

Users can choose between Brief (critical issues only), Standard (alerts and warnings), and Detailed (full analysis with improvement recommendations).

Can I use this with any Claude Code agent?

While optimized for the Task Orchestrator repository, it can validate any skill or subagent that has a documented SKILL.md or pattern definition.

Does this skill block my workflow if an issue is found?

No, Orchestration QA is designed as a non-blocking observer. it reports issues using severity levels like ALERT and WARN but lets the user decide how to proceed.

What kind of deviations does it detect?

It identifies critical violations such as status bypasses, cross-domain task mixing, missing requirements, and suboptimal tool selection.

Orchestration QA

Name: Orchestration QA
Author: jpicklyk

byjpicklyk

•

128

•

Security & Testing

Validates AI orchestration workflows to ensure adherence to documented patterns and provides continuous quality improvement suggestions.

Orchestration QA is a specialized quality control skill designed for the Task Orchestrator ecosystem. It monitors interactions between Claude and specialized subagents to ensure workflows follow established best practices and architectural patterns. By allowing users to interactively enable specific analysis modules—such as token efficiency, execution graph quality, and routing validation—it provides deep insights into AI performance without unnecessary token overhead. It captures pre-execution context and performs post-execution reviews to detect deviations, suggest optimizations, and track recurring patterns for long-term improvement.

Key Features

01Interactive configuration for cost-effective analysis selection

02128 GitHub stars

03Structured deviation reporting with ALERT/WARN/INFO severity levels

04Token efficiency and parallelization opportunity detection

05Pre-execution context capture and checkpoint setting

06Post-execution workflow adherence and output validation

Use Cases

01Optimizing token consumption by identifying redundant tool calls and verbose output patterns.

02Ensuring specialized development subagents follow mandatory lifecycle and testing steps.

03Auditing complex software feature orchestration to prevent process bypass and context loss.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jpicklyk/task-orchestrator orchestration-qa

For use in Claude.ai and ChatGPT

Download Skill