aitbc/.windsurf/skills/openclaw-performance-optimizer.md

description, title, version

description	title	version
Atomic Hermes agent performance tuning and optimization with deterministic outputs	hermes-performance-optimizer	1.1

Hermes Performance Optimizer

Purpose

Optimize agent performance, tune execution parameters, and improve efficiency for Hermes agents through systematic analysis and adjustment.

Activation

Trigger when user requests performance optimization: agent tuning, parameter adjustment, efficiency improvements, or performance benchmarking.

Input

{
  "operation": "tune|benchmark|optimize|profile",
  "agent": "agent_name",
  "target": "speed|memory|throughput|latency|all",
  "parameters": {
    "max_tokens": "number (optional)",
    "temperature": "number (optional)",
    "timeout": "number (optional)"
  }
}

Output

{
  "summary": "Agent performance optimization completed successfully",
  "operation": "tune|benchmark|optimize|profile",
  "agent": "agent_name",
  "target": "speed|memory|throughput|latency|all",
  "before_metrics": {
    "execution_time": "number",
    "memory_usage": "number",
    "throughput": "number",
    "latency": "number"
  },
  "after_metrics": {
    "execution_time": "number",
    "memory_usage": "number",
    "throughput": "number",
    "latency": "number"
  },
  "improvement": {
    "speed": "percentage",
    "memory": "percentage",
    "throughput": "percentage",
    "latency": "percentage"
  },
  "issues": [],
  "recommendations": [],
  "confidence": 1.0,
  "execution_time": "number",
  "validation_status": "success|partial|failed"
}

Process

1. Analyze

• Profile current agent performance
• Identify bottlenecks
• Assess optimization opportunities
• Validate agent state

2. Plan

• Select optimization strategy
• Define parameter adjustments
• Set performance targets
• Plan validation approach

3. Execute

• Apply parameter adjustments
• Run performance benchmarks
• Measure improvements
• Validate stability

4. Validate

• Verify performance gains
• Check for regressions
• Validate parameter stability
• Confirm agent functionality

Constraints

• MUST NOT modify agent core functionality
• MUST NOT exceed 90 seconds for optimization
• MUST validate parameter ranges
• MUST preserve agent behavior
• MUST rollback on critical failures

Environment Assumptions

• Agent operational and accessible
• Performance monitoring available
• Parameter configuration accessible
• Benchmarking tools available
• Agent state persistence functional

Error Handling

• Parameter validation failure → Revert to previous parameters
• Performance regression → Rollback optimization
• Agent instability → Restore baseline configuration
• Timeout during optimization → Return partial results

Example Usage Prompt

Optimize main agent for speed and memory efficiency

Expected Output Example

{
  "summary": "Main agent optimized for speed and memory efficiency",
  "operation": "optimize",
  "agent": "main",
  "target": "all",
  "before_metrics": {
    "execution_time": 15.2,
    "memory_usage": 250,
    "throughput": 8.5,
    "latency": 2.1
  },
  "after_metrics": {
    "execution_time": 11.8,
    "memory_usage": 180,
    "throughput": 12.3,
    "latency": 1.5
  },
  "improvement": {
    "speed": "22%",
    "memory": "28%",
    "throughput": "45%",
    "latency": "29%"
  },
  "issues": [],
  "recommendations": ["Consider further optimization for memory-intensive tasks"],
  "confidence": 1.0,
  "execution_time": 35.7,
  "validation_status": "success"
}

Model Routing Suggestion

Reasoning Model (Claude Sonnet, GPT-4)

• Complex parameter optimization
• Performance analysis and tuning
• Benchmark interpretation
• Regression detection

Performance Notes

• Execution Time: 20-60 seconds for optimization, 5-15 seconds for benchmarking
• Memory Usage: <200MB for optimization operations
• Network Requirements: Agent communication for profiling
• Concurrency: Safe for sequential optimization of different agents