context-engineering

Agent context management for code review involves loading project-specific rules, architecture docs, and source files into the session. This ensures the agent has the necessary information to make accurate suggestions and maintain codebase standards.

Loading project-specific rules and documentation requires parsing a rules file and integrating relevant specs into the session. This provides the agent with the necessary project context to follow best practices during development.

Optimizing AI agent context for debugging integrates source files, error outputs, and test results into the session. This enhances agent understanding of the issue and improves the accuracy of its debugging suggestions.

Conversation history management is supported to maintain relevance and avoid repetition during codebase maintenance. It ensures the agent retains necessary context throughout extended development sessions.

Context management applies directly to security audits, feature development, and codebase maintenance. Deliberately curating the agent context based on project requirements ensures accurate task performance across these development tasks.

Starting a coding session with optimized context requires running the command `context-engineering start-session --project <project-name>`. This initializes the session and loads the necessary project context.