orch-subject-pipeline

okay, here's your JSON: {}

Updated Aug 27, 2026

Search Tags:#evaluation #pipeline #workspaces #orchestrator #grading #start.md #phase-mask

Ready to run

npx skills add https://github.com/AKCqhzdy/dse-subject-grading --skill orch-subject-pipeline

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: orch-subject-pipeline
Source: https://github.com/AKCqhzdy/dse-subject-grading/tree/main/skills-v3/orch-subject-pipeline
Command: npx skills add https://github.com/AKCqhzdy/dse-subject-grading --skill orch-subject-pipeline

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

End-to-end pipeline orchestration for grading and evaluation, enabling a single START.md to drive both sub-orchestrators (orch-subject-grading and orch-subject-evaluation) and produce a combined result without manual hand-wiring.

Core Features & Use Cases

Unified run: orchestrates orch-subject-grading and orch-subject-evaluation from a single START.md.
Phase-mask aware: applies grading/evaluation phase masks and ensures grading output materializes before evaluation begins.
Copy & workspace handling: copies grading outputs between workspaces when needed; supports single-workspace and cross-workspace pipelines.
Deterministic context: reads START.md once and inherits scope/options/phase-selection for both sub-orchestrators.
Observability: emits a combined pipeline manifest with a clear summary of grading, copy steps, and evaluation.

Quick Start

Run this skill with START.md to execute grading and evaluation in a single end-to-end run.

Frequently Asked Questions about orch-subject-pipeline

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

FAQPage Schema

How do I automate end-to-end grading and evaluation pipelines from a single start file?▼

End-to-end grading and evaluation pipelines can be automated from a single START.md file using an orchestrator that drives both sub-orchestrators sequentially. It reads the file once to inherit scope and options, applying phase masks to coordinate grading before evaluation automatically.

What is a phase-mask model in subject grading and evaluation?▼

A phase-mask model in subject grading and evaluation controls pipeline execution phases by applying masks that ensure grading outputs fully materialize before the evaluation sub-orchestrator begins, preventing incomplete data processing.

Do I need separate sub-orchestrators to run the subject grading pipeline?▼

Yes, running the subject grading pipeline requires two separate sub-orchestrators: orch-subject-grading and orch-subject-evaluation. The pipeline orchestrator coordinates these components to produce a combined result without manual hand-wiring.

Can I copy grading outputs between workspaces during the evaluation pipeline?▼

Yes, you can copy grading outputs between workspaces during the evaluation pipeline. The orchestrator supports cross-workspace pipelines, automatically copying materialized grading outputs to the evaluation workspace when needed.

How do I track the status of grading and evaluation steps in an automated pipeline?▼

To track the status of grading and evaluation steps, the automated pipeline emits a combined pipeline manifest. This observability feature provides a clear summary of grading execution, copy steps, and evaluation results.