[IMPORTANT] Use TaskCreate to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ask user whether to skip.
Prerequisites: MUST READ .claude/skills/shared/understand-code-first-protocol.md AND .claude/skills/shared/evidence-based-reasoning-protocol.md before executing.
- docs/project-reference/domain-entities-reference.md — Domain entity catalog, relationships, cross-service sync (read when task involves business entities/models)
Evidence Gate: MANDATORY IMPORTANT MUST — every claim, finding, and recommendation requires file:line proof or traced evidence with confidence percentage (>80% to act, <80% must verify first).
External Memory: For complex or lengthy work (research, analysis, scan, review), write intermediate findings and final results to a report file in plans/reports/ — prevents context loss and serves as deliverable.
Quick Summary
Goal: Analyze and optimize performance bottlenecks in database queries, API endpoints, or frontend rendering.
Workflow:
-
Profile — Identify bottlenecks using profiling data or metrics
-
Analyze — Trace hot paths and measure impact
-
Optimize — Apply targeted optimizations with before/after measurements
Key Rules:
-
Analysis Mindset: measure before and after, never optimize blindly
-
Evidence-based: every claim needs profiling data or benchmarks
-
Focus on highest-impact bottlenecks first
$ARGUMENTS
Analysis Mindset (NON-NEGOTIABLE)
Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).
-
Do NOT assume a bottleneck location — verify with actual code traces and profiling evidence
-
Every performance claim must include file:line evidence
-
If you cannot prove a bottleneck with a code trace, state "suspected, not confirmed"
-
Question assumptions: "Is this really slow?" → trace the actual execution path and query plan
-
Challenge completeness: "Are there other bottlenecks?" → check the full request pipeline
-
No "should improve performance" without proof — measure before and after
[IMPORTANT] Database Performance Protocol (MANDATORY):
-
Paging Required — ALL list/collection queries MUST use pagination. NEVER load all records into memory. Verify: no unbounded GetAll() , ToList() , or Find() without Skip/Take or cursor-based paging.
-
Index Required — ALL query filter fields, foreign keys, and sort columns MUST have database indexes configured. Verify: entity expressions match index field order, database collections have index management methods, migrations include indexes for WHERE/JOIN/ORDER BY columns.
⚠️ MANDATORY: Confidence & Evidence Gate
MANDATORY IMPORTANT MUST declare Confidence: X% with profiling data + file:line proof for EVERY claim. 95%+ recommend freely | 80-94% with caveats | 60-79% list unknowns | <60% STOP — gather more evidence.
Activate arch-performance-optimization skill and follow its workflow.
CRITICAL: Present findings and optimization plan. Wait for explicit user approval before making changes.
IMPORTANT Task Planning Notes (MUST FOLLOW)
-
Always plan and break work into many small todo tasks
-
Always add a final review todo task to verify work quality and identify fixes/enhancements
Workflow Recommendation
IMPORTANT MUST: If you are NOT already in a workflow, use AskUserQuestion to ask the user:
-
Activate quality-audit workflow (Recommended) — performance → sre-review → test
-
Execute /performance directly — run this skill standalone
Next Steps
MANDATORY IMPORTANT MUST after completing this skill, use AskUserQuestion to recommend:
-
"/sre-review (Recommended)" — Production readiness review after optimization
-
"/changelog" — Document performance changes
-
"Skip, continue manually" — user decides
Closing Reminders
MANDATORY IMPORTANT MUST break work into small todo tasks using TaskCreate BEFORE starting. MANDATORY IMPORTANT MUST validate decisions with user via AskUserQuestion — never auto-decide. MANDATORY IMPORTANT MUST add a final review todo task to verify work quality.