context-engineering-collection

Context engineering standardizes AI agent context management by enforcing progressive disclosure, memory architectures, and disciplined tool integration to prevent fragmentation across platforms like Cursor and Claude Code.

Progressive disclosure in multi-agent orchestration loads metadata at startup and full instructions on activation, optimizing memory systems and preventing context degradation across distributed agent workflows.

AI agent evaluation uses LLM-as-a-judge patterns and production-grade evaluation frameworks to assess context compression, multi-agent orchestration, and tool design effectiveness within development workflows.

This context engineering collection is platform-agnostic and supports Cursor, Claude Code, and other agent platforms by providing foundational patterns, filesystem-based context, and cross-skill references.

Context compression for AI agents should be used when managing large memory systems and multi-agent patterns, preventing context degradation and optimizing operational tool design in production environments.

Filesystem-based context for AI agents should follow architectural patterns with optional scripts, templates, and cross-skill references that standardize agent design and enforce progressive disclosure.