/reflect

Reflect on this conversation session and extract key learnings to save as context files.

Analysis Focus

Analyze the current session for:

1. Successes - Effective approaches, good decisions, what worked well
2. Failures - What didn't work, challenges, mistakes (valuable learning!)
3. Unknowns - Knowledge gaps discovered, things we didn't know
4. Key Learnings - Actionable insights for future work

$ARGUMENTS

Output Structure

Create a structured reflection:

# Session Reflection - [Brief Description]

## Successes
- [List successes with context]

## Failures & Lessons
- [List failures with what was learned]

## Unknowns Discovered
- [List knowledge gaps and why they matter]

## Key Learnings
- [Actionable insights for future work]

Process Learnings

After creating the reflection, invoke the context-manager agent via Task tool:

Prompt: "Process this session reflection and create appropriate context files (c₁-c₅):

[Include full reflection]

Autonomously classify learnings:
- c₁: Methodology improvements (HOW to do things better)
- c₂: New domain knowledge (WHAT we learned)
- c₃: Tool capabilities discovered
- c₄: Historical lessons from this session
- c₅: State or configuration updates

Create/update context files with semantic versioning. Focus on actionable, specific learnings."

Output

Summarize:

• Context files created/updated
• Which components (c₁-c₅) affected
• Key learnings captured
• How these will help future work

Note: Skip reflection if session was trivial (<5 tool calls, no significant tasks).

Usage Examples

Example 1: General Session Reflection

/reflect

Analyzes the entire session and extracts all learnings across all categories.

Example 2: Focus on Specific Area

/reflect Focus on the authentication implementation approach

Concentrates reflection on authentication-related decisions and learnings.

/reflect deployment workflow

Analyzes only deployment-related successes, failures, and learnings.

Example 3: After Complex Implementation

User completes: "Implement OAuth2 authentication with JWT tokens"
User runs: /reflect

Expected output:
- c₁ context: Authentication workflow methodology
- c₂ context: OAuth2 vs JWT trade-offs knowledge
- c₄ context: Lessons from initial failed approach with session cookies

Example 4: After Debugging Session

User completes: "Debug and fix race condition in user registration"
User runs: /reflect race condition debugging

Expected output:
- c₁ context: Debugging methodology for async issues
- c₂ context: Race condition patterns in Node.js
- c₄ context: Specific fix that worked (mutex implementation)

Example 5: After Research/Discovery

User completes: "Research best practices for API versioning"
User runs: /reflect

Expected output:
- c₂ context: API versioning strategies comparison
- c₄ context: Decision rationale for choosing semantic versioning

When to Use

Use /reflect after:

• Complex implementations (multi-step features)
• Problem-solving sessions (debugging, troubleshooting)
• Research and discovery work
• Failed attempts that led to insights
• Any session where you learned something valuable

Skip /reflect for:

• Simple file reads or edits
• Trivial command executions
• Sessions with no significant learnings