Deep Planning Command
/deep-planning converts Sypha into a thorough architect who examines your codebase, poses clarifying questions, and develops a detailed implementation plan prior to writing any code.

When you employ /deep-planning, Sypha executes a four-phase process that reflects how experienced developers tackle complex features: comprehensive investigation, discussion & clarification of requirements, meticulous planning, and organized task creation with progress monitoring.
The Four-Step Process
Step 1: Silent Investigation
Sypha assumes the role of a detective, quietly exploring your codebase to comprehend its architecture, patterns, and limitations. He inspects source files, evaluates import patterns, uncovers class hierarchies, and detects technical debt indicators. No commentary, no narration - purely concentrated research.
Throughout this phase, Sypha executes commands such as:
- Locating all class and function definitions throughout your codebase
- Evaluating import patterns to comprehend dependencies
- Uncovering project structure and file arrangement
- Detecting TODOs and technical debt
Step 2: Discussion and Questions
After Sypha comprehends your codebase, he poses focused questions that will influence the implementation. These aren't broad questions - they're tailored to your project and the feature you're developing.
Questions may address:
- Clarifying unclear requirements
- Selecting between equally viable implementation strategies
- Validating assumptions about system functionality
- Understanding preferences for technical choices
Step 3: Implementation Plan Document
Sypha produces a structured markdown document (implementation_plan.md) that functions as your implementation blueprint. This isn't an ambiguous outline - it's a comprehensive specification with precise file paths, function signatures, and implementation sequence.
The plan encompasses eight thorough sections:
- Overview: The objective and high-level strategy
- Types: Complete type definitions and data structures
- Files: Precise files to create, modify, or delete
- Functions: New and modified functions with signatures
- Classes: Class modifications and inheritance specifics
- Dependencies: Package requirements and versions
- Testing: Validation approaches and test requirements
- Implementation Order: Sequential execution steps
Step 4: Implementation Task Creation
Sypha generates a new task that references the plan document and incorporates trackable implementation steps. The task arrives with particular commands to access each section of the plan, guaranteeing the implementing agent (whether that's you or Sypha in Act Mode) can navigate the blueprint effectively.
Deep Planning integrates seamlessly with Focus Chain. The implementation steps automatically transform into a todo list with real-time progress monitoring, maintaining complex projects organized and on course.
Using Deep Planning
Initiate a deep planning session by entering /deep-planning followed by your feature description:
/deep-planning Add user authentication with JWT tokens and role-based access controlSypha will commence his investigation promptly. You'll observe him reading files and executing commands to understand your codebase. After he's collected sufficient context, he'll engage you in discussion prior to generating the plan.
Example Workflow
Here's how I utilize /deep-planning for an actual feature:
I enter /deep-planning implement a caching layer for API responses
Sypha explores my codebase, analyzing:
- Current API architecture and endpoints
- Existing data flow patterns
- Database queries and performance constraints
- Configuration and environment configuration
Sypha inquires:
- "Should we utilize Redis or in-memory caching?"
- "What's the acceptable cache staleness for user data?"
- "Do you require cache invalidation webhooks?"
Sypha produces implementation_plan.md containing:
- Cache service class specifications
- Redis connection configuration
- Modified API endpoints with caching logic
- Cache key generation approaches
- TTL configurations for various data types
Sypha establishes a new task with:
- Reference to the implementation plan
- Commands to access specific sections
- Trackable todo items for each implementation step
- Request to transition to Act Mode for execution
Integration with Plan/Act Mode
Deep Planning is engineered to integrate seamlessly with Plan/Act Mode:
- Utilize
/deep-planningin Plan Mode for the investigation and planning phases - The generated task requests transitioning to Act Mode for implementation
- Focus Chain automatically monitors progress through the implementation steps
This separation guarantees planning remains concentrated on architecture while implementation remains concentrated on execution.
Best Practices
When to Use Deep Planning
Apply /deep-planning for:
- Features affecting multiple parts of your codebase
- Architectural modifications requiring meticulous coordination
- Complex integrations with external services
- Refactoring initiatives that demand systematic execution
- Any feature where you'd typically spend time whiteboarding
Making the Most of Investigation
Allow Sypha to complete his investigation comprehensively. The plan's quality directly correlates with how thoroughly he understands your codebase. If you have particular areas he should investigate, reference them in your initial request.
Reviewing the Plan
Always examine implementation_plan.md prior to beginning implementation. The plan is thorough but not unchangeable - you can modify it directly if necessary. Consider it as a collaborative document between you and Sypha.
Tracking Progress
With Focus Chain activated, your implementation progress appears in the task header. Each completed step receives automatic verification as Sypha progresses through the plan, providing you real-time visibility into complex implementations.
Inspiration
I utilize /deep-planning whenever I'm preparing to build something that would typically require a design document. Recent examples from my workflow:
-
Migrating authentication systems: Deep Planning charted every endpoint, identified all authentication touchpoints, and developed a migration plan that prevented breaking changes.
-
Adding real-time features: The plan encompassed WebSocket integration, event handling, state synchronization, and fallback mechanisms for disconnections.
-
Database schema refactoring: Sypha identified all impacted queries, generated migration scripts, and strategized the rollout to minimize downtime.
-
API versioning implementation: The plan outlined route changes, backward compatibility layers, deprecation notices, and client migration paths.
The strength of /deep-planning is that it enforces thoughtful architecture prior to implementation. It's like having a senior developer assess your approach before you write code, except that developer possesses complete knowledge of your entire codebase.
Deep Planning demands models with robust reasoning capabilities. It performs optimally with the latest generation of models, such as GPT-5, Claude 4, Gemini 2.5, or Grok 4. Smaller models may have difficulty with the thorough analysis required.
For simpler tasks that don't necessitate extensive planning, consider utilizing /newtask to establish focused tasks with context, or proceed directly into implementation if the path forward is evident.