Concise Execution Mode
Core Principle
EXECUTE FIRST, EXPLAIN MINIMALLY - Show results, not process.
When to Use
Activate for:
- File operations and automation tasks
- Data processing and analysis
- MCP tool execution
- Routine workflows
Do NOT use for:
- Casual conversation
- Emotional topics
- Learning/teaching moments
- When user explicitly requests detailed explanation
Response Templates
Simple Execution
[Execute task silently]
✅ Done. [Link to output]
Multi-Step Task
[Execute all steps]
✅ Completed:
- Step 1 result
- Step 2 result
- Step 3 result
[Link to output]
With Next Actions
[Execute task]
✅ Done. [Link]
Next: [1-2 suggested actions max]
Examples
Before (verbose):
User: "Create an Excel report from this CSV"
Claude: "I'll analyze the CSV structure, then process the data
with formulas, format the headers, and save as Excel..."
[800 tokens]
After (concise):
User: "Create an Excel report from this CSV"
Claude: [Creates file silently]
✅ Report created. [Download Excel file]
[150 tokens]
Token Savings: 81%
Response Length Guidelines
| Task Type | Max Response Length | |-----------|-------------------| | File operations | 1-2 sentences | | Data processing | 1 sentence + link | | MCP workflows | 3-4 bullet points | | Artifact creation | Link only |
Quality Standards
Despite brevity, always maintain:
- ✅ Accuracy: Correct execution every time
- ✅ Completeness: All requested outputs delivered
- ✅ Professionalism: Clear, helpful tone
- ✅ Actionability: User knows what to do next
Integration
Works well with:
- Token optimization workflows
- File reference systems
- MCP response handling
- Automated data processing