/full-context

Intelligently gathers relevant project context using an adaptive sub-agent strategy before proceeding with a user's request.

This is the primary command for starting any significant task, such as implementing a new feature, fixing a complex bug, or planning an architectural change. It ensures the AI has a deep and comprehensive understanding of the current state of the project before taking action.

Auto-Loaded Project Context:

@/CLAUDE.md
@/docs/ai-context/project-structure.md
@/docs/ai-context/docs-overview.md

Command Execution

The /full-context command is adaptive, choosing the most efficient analysis strategy based on the complexity of your request.

Step 1: Intelligent Analysis Strategy Decision

Based on your request (e.g., /full-context "add waveform display to the audio player") and the auto-loaded project context, the AI decides on an optimal approach:

  • Direct Approach (0-1 sub-agents): For simple questions or tasks that can be answered by reading a few specific documentation or code files.
  • Focused Investigation (2-3 sub-agents): For tasks confined to a single domain (e.g., just the frontend) but requiring a deep dive into its dependencies and impacts.
  • Multi-Perspective Analysis (3+ sub-agents): For complex tasks that span multiple components (e.g., frontend and backend), requiring careful dependency mapping and impact assessment across the system.

Step 2: Autonomous Sub-Agent Design

If a sub-agent approach is chosen, the AI has complete freedom to design custom sub-agents tailored to the request. It uses the project's documentation to design agents that can:

  • Analyze specific components or code files.
  • Map import/export dependencies to find all affected files.
  • Assess the impact on tests, configurations, and related components.
  • Verify that any proposed changes align with existing project patterns.
  • Plan for cleanup of any obsolete code that a change might create.

All sub-agents are launched in parallel for maximum efficiency.

Step 3: Execution and Synthesis

The AI collects the findings from all sub-agents and synthesizes them into a single, comprehensive understanding. For tasks involving code changes, this synthesis results in a detailed implementation plan that includes dependency updates, a list of affected files, and necessary cleanup tasks.

Step 4: Consider MCP Server Usage (Optional)

After gathering internal context, the AI may decide to consult an MCP server like Gemini for architectural advice or Context7 for up-to-date library documentation if the task requires it.

Step 5: Context Summary and Implementation Plan

Before proceeding with the implementation, the AI provides a concise (2-4 sentence) status update summarizing its findings and its planned strategy. This gives you a clear picture of what the AI has learned and how it intends to tackle your request.

Example status update:

Analysis revealed the voice pipelines use Socket.IO for real-time communication. I'll implement the new transcription feature by extending the existing Socket.IO event handling in both the FastAPI backend and SvelteKit frontend, following the established pattern. This will require updating 3 import statements and adding exports to the socket handler module.

After this summary, the AI proceeds with the implementation, armed with a deep and validated understanding of your project.