context-engineering

Organizes hierarchical AI context to enforce project-specific conventions and resolve conflicts.

Updated Apr 19, 2026

Search Tags:#workflow #architecture #context #agent #engineering #prompting

Ready to run

npx skills add https://github.com/dmarins/sdd --skill context-engineering-dmarins

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: context-engineering
Source: https://github.com/dmarins/sdd/tree/main/skills/context-engineering
Command: npx skills add https://github.com/dmarins/sdd --skill context-engineering-dmarins

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill solves the common issue of AI agents losing focus, hallucinating, or ignoring project conventions due to poorly structured or excessive context.

Core Features & Use Cases

Context Hierarchy Management: Provides a structured framework to organize information from persistent rules to transient task-specific data.
Conflict Resolution: Offers standardized patterns for handling ambiguous requirements and conflicting instructions between specs and existing code.
Use Case: When starting a new feature in a complex Go/AWS project, use this skill to load only the relevant architecture specs, local reference implementations, and project-wide rules to ensure the agent produces high-quality, compliant code.

Quick Start

Apply the context engineering framework to your current session by summarizing the project rules and loading only the specific files required for the immediate task.

Frequently Asked Questions about context-engineering

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

FAQPage Schema

How do I stop my AI agent from hallucinating and ignoring project conventions?▼

To stop AI agent hallucination and enforce project conventions, you must optimize agent context by loading only specific files required for the immediate task. This structured, hierarchical context management prevents agent drift and reduces token waste.

How do I organize persistent rules and transient task data for AI development sessions?▼

Organize persistent rules and transient task data for AI development sessions using context hierarchy management. This framework structures information delivery from project-wide rules down to task-specific data, ensuring agents maintain focus and adhere to architectural specs.

What is the best way to resolve conflicting instructions between specs and existing code?▼

The best way to resolve conflicting instructions between specs and existing code is applying standardized conflict resolution patterns. These patterns handle ambiguous requirements within the agent context, ensuring precise information delivery and architectural alignment without drift.

How do I load only relevant architecture specs for a complex Go and AWS feature?▼

Load relevant architecture specs for complex Go and AWS features by summarizing project rules and filtering context to specific files required for the immediate task. This precise information delivery ensures the agent produces high-quality, compliant code.

When do I need structured context management for my AI agent?▼

You need structured context management for your AI agent when starting new features in complex projects requiring precise information delivery. It satisfies the need to prevent agent drift, reduce token waste, and enforce adherence to project-specific conventions.