skills.rest
synthesisengineering
synthesisengineeringOFFICIAL
Explore all skills by this author →
WebsiteGitHub

synthesis-context-lifecycle

Implement a tiered context architecture with CONTEXT.md, REFERENCE.md, and sessions/ archives.

15|2|Updated Mar 17, 2026
Search Tags:#context-management#project-memory#ai-collaboration#working-memory#tiered-memory#reference-facts#session-archiving
One-click install
npx skills add https://github.com/synthesisengineering/synthesis-skills --skill synthesis-context-lifecycle-synthesisengineering
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: synthesis-context-lifecycle
Source: https://github.com/synthesisengineering/synthesis-skills/tree/main/synthesis-context-lifecycle
Command: npx skills add https://github.com/synthesisengineering/synthesis-skills --skill synthesis-context-lifecycle-synthesisengineering

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI collaborators start every session with zero context. For long-running projects, context must be tiered and lifecycle-managed to preserve working memory, semantic memory, and episodic history across sessions.

Core Features & Use Cases

  • Three-tier context architecture (CONTEXT.md, REFERENCE.md, sessions/ archive) for scalable memory management.
  • Archival protocol to move completed tasks and stale session data into references/sessions, keeping CONTEXT.md lean.
  • Guidance on project organization and lifecycle transitions to maintain context quality over time.

Quick Start

Initialize a tiered-context workspace for a project to begin capturing active memory, stable reference facts, and episodic session logs.

Frequently Asked Questions about synthesis-context-lifecycle

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

FAQPage Schema
How do I manage AI context and working memory across long-running projects?▼

A tiered context architecture organizes project memory into CONTEXT.md for active working memory, REFERENCE.md for stable facts, and sessions/ for episodic history. This prevents AI collaborators from starting with zero memory.

How do I keep AI project memory files from growing too large over multiple sessions?▼

Apply a formal archival protocol to move completed tasks and stale session data from CONTEXT.md into REFERENCE.md or the sessions/ archive. This ensures CONTEXT.md stays under 150 lines.

What is the best way to preserve AI session history for knowledge-heavy workflows?▼

Maintain a sessions/ directory for episodic history alongside tiered context files. This captures episodic logs while stable facts are moved to REFERENCE.md.

When should I move stable facts out of my AI project's active context?▼

Move stable facts out of active context during project lifecycle transitions by migrating them to REFERENCE.md. This keeps active working memory lean and preserves semantic memory without losing reference data.

Does tiered context management work for research and development workflows?▼

Yes, tiered context management works for research, development, and knowledge-heavy workflows. It applies a tiered layout and archival protocol to preserve working memory, semantic memory, and episodic history across many AI interactions.

Related Skills

Explore More Product & Management Skills→
kiryteokiryteo
1

memory-management

Decode shorthand, acronyms, and nicknames into structured working memory for Claude.

Community
Advanced
jinglemansweepjinglemansweep

jp-summary

Aggregate phase artifacts into a structured Phase Summary document.

Community
Advanced
handsomelong922handsomelong922

self-improvement

Log learnings and corrections to .learnings markdown files for future sessions.

Community
Advanced
oakossoakoss
14

agent-standards

Define behavioral and cognitive standards for AI engineering agents.

Official
Advanced