Skip to content
Development
Command

/metrics

Show agent performance metrics

From plugin
claude-flow
67k194 skills157 agents194 commands1 MCP
Install
> /plugin marketplace add ruvnet/claude-flow

How it fires

How this command gets triggered: by you, by Claude, or both.

  • Fires itselfClaude auto-loads it when your prompt matches the work.
  • You can call itInvoke it directly when you want it.
  • Slash command/metrics

Context preview

What this command does when you run it.

Show agent performance metrics

Command definition

metrics.md
name: metrics
description: Show agent performance metrics
type: command

Agent Metrics Command

Display comprehensive performance metrics for agents including V3 performance gains.

Usage

npx claude-flow agent metrics [agent-id] [options]

Options

| Option | Short | Description | Default | |--------|-------|-------------|---------| | `--period` | `-p` | Time period (1h, 24h, 7d, 30d) | 24h | | `--format` | | Output format (table, json) | table |

Examples

# Overall metrics for last 24 hours
npx claude-flow agent metrics

# Metrics for specific agent
npx claude-flow agent metrics coder-lx7m9k2

# Last hour
npx claude-flow agent metrics -p 1h

# Last 7 days
npx claude-flow agent metrics --period 7d

# JSON output
npx claude-flow agent metrics --format json

Output

Agent Metrics (24h)

+--------------------+----------+
| Metric             |    Value |
+--------------------+----------+
| Total Agents       |        4 |
| Active Agents      |        3 |
| Tasks Completed    |      127 |
| Success Rate       |    96.2% |
| Total Tokens       | 1,234,567|
| Avg Response Time  |    1.45s |
+--------------------+----------+

By Agent Type
+------------+-------+-------+---------+
| Type       | Count | Tasks | Success |
+------------+-------+-------+---------+
| coder      |     2 |    45 |     97% |
| researcher |     1 |    32 |     95% |
| tester     |     1 |    50 |     98% |
+------------+-------+-------+---------+

V3 Performance Gains
  - Flash Attention: 2.8x speedup
  - Memory Reduction: 52%
  - Search: 150x faster

Metrics Explained

Summary Metrics

| Metric | Description | |--------|-------------| | Total Agents | All agents spawned in period | | Active Agents | Currently running agents | | Tasks Completed | Successfully completed tasks | | Success Rate | Percentage of successful tasks | | Total Tokens | Token usage across all agents | | Avg Response Time | Mean task completion time |

V3 Performance Gains

| Metric | Target | Description | |--------|--------|-------------| | Flash Attention | 2.49x-7.47x | Neural attention speedup | | Memory Reduction | 50-75% | Quantization savings | | HNSW Search | 150x-12,500x | Vector search improvement | | SONA Adaptation | <0.05ms | Real-time learning |

JSON Output

{
  "period": "24h",
  "summary": {
    "totalAgents": 4,
    "activeAgents": 3,
    "tasksCompleted": 127,
    "avgSuccessRate": "96.2%",
    "totalTokens": 1234567,
    "avgResponseTime": "1.45s"
  },
  "byType": [
    { "type": "coder", "count": 2, "tasks": 45, "successRate": "97%" }
  ],
  "performance": {
    "flashAttention": "2.8x speedup",
    "memoryReduction": "52%",
    "searchImprovement": "150x faster"
  }
}

Related Commands

  • `npx claude-flow agent status` - Individual agent metrics
  • `npx claude-flow performance benchmark` - Full performance suite
  • `npx claude-flow status` - System-wide status
Read more
Ships withclaude-flow

An agent meta-harness for Claude Code and Codex. Agent = Model + Harness. The model writes; the harness gives it tools, memory, loops, sandboxes, and controls so it can actually work.

Get the whole plugin