How does it track workflow progress?

It generates a session-meta.json file that tracks the session ID, workflow type, start/end times, completion status, and a mapping of all generated artifacts.

What problem does Session Isolation solve?

It prevents file collisions where sequential or concurrent AI tasks overwrite files with identical names, such as design.md or report.md, by isolating them in unique directories.

Does this require changing my existing agents?

No, the pattern includes a backward-compatible legacy mode. Agents will function normally using default paths if a session path is not explicitly provided.

Can I use this with multiple AI models?

Yes, the metadata schema is designed to track which specific models (e.g., Claude, Grok, Gemini) were used for different phases of the session.

Session Isolation Pattern

Name: Session Isolation Pattern
Author: MadAppGang

byMadAppGang

•

196

생산성 및 워크플로

Organizes multi-artifact workflows into isolated session directories to prevent file collisions and preserve version history.

소개

The Session Isolation Pattern provides a robust framework for managing complex AI-driven workflows that generate multiple documents, such as design specs, code reviews, and reports. By dynamically creating unique session IDs and folder structures, it ensures that concurrent or sequential tasks do not overwrite each other's work. This skill is essential for developers using Claude for multi-stage processes where tracking metadata and maintaining distinct artifact sets are critical for project integrity and auditability.

주요 기능

Graceful fallback to legacy paths for backward compatibility
Context-aware artifact path mapping for sub-agent orchestration
Structured review folder hierarchy for multi-phase auditing
196 GitHub stars
JSON metadata tracking for workflow status and model usage
Automatic unique session directory generation with timestamps and slugs

사용 사례

Running concurrent SEO or design reviews on the same codebase without data loss
Archiving historical versions of AI-generated project plans and reports
Orchestrating multi-agent workflows that require distinct output paths

소개

주요 기능

Graceful fallback to legacy paths for backward compatibility
Context-aware artifact path mapping for sub-agent orchestration
Structured review folder hierarchy for multi-phase auditing
196 GitHub stars
JSON metadata tracking for workflow status and model usage
Automatic unique session directory generation with timestamps and slugs

사용 사례

Running concurrent SEO or design reviews on the same codebase without data loss
Archiving historical versions of AI-generated project plans and reports
Orchestrating multi-agent workflows that require distinct output paths