capability-evolver

name	Capability Evolver
description	Meta-skill for AI agent self-improvement. Analyzes runtime logs to detect error patterns, regressions, and inefficiencies, then generates structured improvement proposals. Use when the user or agent asks to analyze logs, diagnose failures, improve agent reliability, generate evolution proposals, or assess system health. Supports analyze, evolve, and status actions.

Permission Setup Review

python scripts/_prepare_workspace.py is an optional setup helper for the Capability Evolver skill. It may apply broad filesystem permissions for compatibility. Show the target paths and permission mode, explain the risk, and prefer a narrower mode when shared write access is not needed.

Capability Evolver

Local skill by Claw0x — runs entirely in your OpenClaw agent.

Runs locally. No external API calls, no API key required. Complete privacy.

Analyze agent runtime logs, detect patterns, compute health scores, and generate structured improvement proposals. Pure deterministic logic — no LLM, no external dependencies.

Quick Reference

When This Happens	Use Action	What You Get
Agent keeps failing	analyze	Error patterns + health score
Same error repeats	analyze	Root cause identification
Need improvement plan	evolve	Prioritized recommendations
System health check	status	Health score + summary
Post-deployment review	analyze	Regression detection
Fleet-wide diagnostics	analyze (batch)	Cross-agent patterns

Why deterministic? Reproducible results, no hallucination risk, sub-100ms processing, zero token costs.

---

Prerequisites

None. Just install and use.

5-Minute Quickstart

Step 1: Install (30 seconds)

openclaw skill add capability-evolver

Step 2: Analyze Your First Logs (1 minute)

const result = await agent.run('capability-evolver', {
  action: 'analyze',
  logs: [
    {timestamp: '2025-01-15T10:00:00Z', level: 'error', message: 'ETIMEDOUT', context: 'payment-api.ts'},
    {timestamp: '2025-01-15T10:01:00Z', level: 'error', message: 'ETIMEDOUT', context: 'payment-api.ts'},
    {timestamp: '2025-01-15T10:02:00Z', level: 'error', message: 'ETIMEDOUT', context: 'payment-api.ts'}
  ]
});

Step 3: Get Actionable Insights (instant)

{
  "patterns": [
    {
      "type": "repeated_error",
      "severity": "high",
      "description": "ETIMEDOUT appeared 3 times in payment-api.ts",
      "affected_contexts": ["payment-api.ts"]
    }
  ],
  "health_score": 45,
  "recommendations": [
    "Add timeout configuration to payment-api.ts",
    "Implement retry logic with exponential backoff",
    "Monitor payment API response times"
  ]
}

Step 3: Generate Evolution Plan (instant)

const evolution = await agent.run('capability-evolver', {
  action: 'evolve',
  logs: result.logs,
  strategy: 'harden'
});

Done. You now have a prioritized improvement roadmap, all processed locally.

---

Real-World Use Cases

Scenario 1: Production Incident Response

Problem: Your agent crashed in production and you need to understand why

Solution:

1. Export last 1000 log entries
2. Run analyze action
3. Get error patterns and cascades
4. Identify root cause in minutes

Example:

const logs = await db.logs.findMany({ 
  where: { timestamp: { gte: incidentStart } },
  orderBy: { timestamp: 'asc' }
});

const analysis = await agent.run('capability-evolver', {
  action: 'analyze',
  logs: logs.map(l => ({
    timestamp: l.timestamp,
    level: l.level,
    message: l.message,
    context: l.context
  }))
});

// analysis.patterns shows: "auth-service.ts failed, then payment-api.ts failed"
// Root cause: auth service timeout cascaded to payment failures

Scenario 2: Continuous Improvement Pipeline

Problem: You want your agent to automatically improve based on production data

Solution:

1. Schedule daily log analysis
2. Generate evolution proposals
3. Review and apply recommendations
4. Track improvement over time

Example:

// Cron job: every day at 2am
async function dailyEvolution() {
  const logs = await getLast24HoursLogs();
  
  const evolution = await agent.run('capability-evolver', {
    action: 'evolve',
    logs,
    strategy: 'balanced'
  });
  
  // Store recommendations for review
  for (const rec of evolution.recommendations.filter(r => r.priority === 'critical')) {
    await db.recommendations.create({
      title: `${rec.category}: ${rec.description}`,
      priority: rec.priority,
      affected_files: rec.affected_files,
      approach: rec.suggested_approach
    });
  }
  
  // Track health score trend
  await db.metrics.create({
    date: new Date(),
    health_score: evolution.estimated_improvement
  });
}
// Result: Health score improved from 45 to 85 over 3 months

Scenario 3: Multi-Agent Fleet Management

Problem: Managing 50+ agent instances, need to identify systemic issues

Solution:

1. Aggregate logs from all agents
2. Batch analyze to find common patterns
3. Fix once, deploy to all agents
4. Reduce fleet-wide error rate

Example:

# Collect logs from all agents
all_logs = []
for agent_id in agent_fleet:
    logs = fetch_agent_logs(agent_id, last_24h)
    all_logs.extend(logs)

# Analyze fleet-wide
result = client.call("capability-evolver", {
    "action": "analyze",
    "logs": all_logs
})

# result.patterns shows: "40 of 50 agents failing on auth-service.ts"
# Fix auth-service.ts once, deploy to all agents
# Result: 80% reduction in fleet-wide errors

Scenario 4: Pre-Deployment Health Check

Problem: Want to ensure new deployment doesn't introduce regressions

Solution:

1. Analyze logs from staging environment
2. Compare health score to production baseline
3. Block deployment if health score drops
4. Catch regressions before production

Example:

// Pre-deployment health check script
async function preDeploymentCheck() {
  const stagingLogs = await fetchStagingLogs();
  
  const result = await agent.run('capability-evolver', {
    action: 'analyze',
    logs: stagingLogs
  });
  
  const BASELINE = 75;
  
  if (result.health_score < BASELINE) {
    console.error(`Health score ${result.health_score} below baseline ${BASELINE}`);
    console.error('Critical patterns:', result.patterns.filter(p => p.severity === 'critical'));
    process.exit(1);
  }
  
  console.log(`✓ Health check passed: ${result.health_score}`);
}
// Result: Zero regression-related incidents in 6 months

---

Integration Recipes

OpenClaw Agent

// Analyze logs after each run
agent.onComplete(async () => {
  const logs = agent.getRecentLogs();
  
  const analysis = await agent.run('capability-evolver', {
    action: 'analyze',
    logs
  });
  
  if (analysis.health_score < 70) {
    console.warn('⚠️ Health score low:', analysis.health_score);
    console.log('Recommendations:', analysis.recommendations);
  }
});

LangChain Agent

def analyze_agent_health(logs):
    result = agent.run("capability-evolver", {
        "action": "analyze",
        "logs": logs
    })
    
    return {
        "health_score": result["health_score"],
        "patterns": result["patterns"],
        "recommendations": result["recommendations"]
    }

# Use in monitoring
health = analyze_agent_health(agent.logs)
if health["health_score"] < 70:
    alert_team(health)

Custom Monitoring Dashboard

// Real-time health monitoring
async function updateHealthDashboard() {
  const logs = await db.logs.findMany({
    where: { timestamp: { gte: Date.now() - 3600000 } } // last hour
  });
  
  const result = await agent.run('capability-evolver', {
    action: 'analyze',
    logs
  });
  
  // Update dashboard
  dashboard.update({
    healthScore: result.health_score,
    errorRate: result.summary.error_count / result.summary.total_logs,
    topPatterns: result.patterns.slice(0, 5)
  });
}

setInterval(updateHealthDashboard, 60000); // every minute

Evolution Strategy Comparison

// Compare different evolution strategies
const logs = await getProductionLogs();

const strategies = ['balanced', 'innovate', 'harden', 'repair-only'];

const results = await Promise.all(
  strategies.map(strategy =>
    agent.run('capability-evolver', {
      action: 'evolve',
      logs,
      strategy
    })
  )
);

// Compare estimated improvements
for (let i = 0; i < strategies.length; i++) {
  console.log(`${strategies[i]}: ${results[i].estimated_improvement}`);
}

// Choose best strategy for current situation
const best = results.reduce((a, b) => 
  parseFloat(a.estimated_improvement) > parseFloat(b.estimated_improvement) ? a : b
);

---

How It Works — Under the Hood

Capability Evolver is a deterministic analysis engine that processes structured log data and produces actionable diagnostics. No LLM is involved �?the analysis is rule-based, which means results are reproducible and fast.

Analysis Engine

The core engine processes log entries through several analysis passes:

1. Pattern detection �?logs are grouped by context (file/module) and level (error/warn/info/debug). The engine looks for:

   • Repeated errors �?the same error message appearing multiple times indicates a systemic issue, not a transient failure
   • Error cascades �?errors in module A followed by errors in module B within a short time window suggest a dependency chain failure
   • Regression signals �?errors that appear after a period of clean logs suggest a recent change broke something
   • Inefficiency patterns �?excessive warn-level logs or repeated retries indicate performance issues

2. Health scoring �?a system health score (0�?00) is computed based on:

   • Error rate (errors / total logs)
   • Error diversity (unique error messages / total errors)
   • Warn-to-error ratio
   • Time distribution (clustered errors score worse than spread-out errors)

3. Recommendation generation �?based on detected patterns, the engine generates specific, actionable recommendations. These aren't generic advice �?they reference the actual files, error messages, and patterns found in your logs.

Evolution Strategies

When using the evolve action, you can choose a strategy that shapes the recommendations:

Strategy	Focus	Best For
auto	Balanced based on health score	Default �?let the engine decide
balanced	Equal weight to reliability and features	Stable systems with moderate issues
innovate	Prioritize new capabilities	Healthy systems ready to grow
harden	Prioritize reliability and error reduction	Systems with frequent failures
repair-only	Fix critical issues only	Systems in crisis

Evolution Proposals

The evolve action produces structured improvement proposals with:

• A unique evolution_id for tracking
• Prioritized recommendations with category labels (reliability, performance, architecture)
• Risk assessment (how risky is each proposed change)
• Estimated improvement (projected health score after implementing recommendations)

Why Deterministic (Not LLM)?

• Reproducible �?same logs always produce the same analysis. Critical for debugging and auditing.
• Fast �?sub-100ms processing. No API call to an AI provider.
• No hallucination risk �?the engine only reports patterns it actually found in the data.
• Cost-effective �?pure computation, no token costs.

The tradeoff: the engine can't understand semantic meaning in log messages the way an LLM could. It relies on structural patterns (frequency, timing, severity) rather than understanding what the error message means in context.

About Claw0x

This skill is provided by Claw0x, the native skills layer for AI agents.

Cloud version available: For users who need centralized analytics and cross-agent insights, a cloud version is available at claw0x.com/skills/capability-evolver.

Explore more skills: claw0x.com/skills

GitHub: github.com/kennyzir/capability-evolver

When to Use

• User says "analyze these logs", "what's failing", "improve my agent", "check system health"
• Agent pipeline needs automated diagnostics after a run
• User wants structured recommendations for fixing recurring errors
• Building a self-healing agent that adapts based on its own failure patterns

Input

Field	Type	Required	Description
input.action	string	yes	"analyze", "evolve", or "status"
input.logs	array	yes (for analyze/evolve)	Array of log entries
input.logs[].timestamp	string	yes	ISO timestamp
input.logs[].level	string	yes	"error", "warn", "info", or "debug"
input.logs[].message	string	yes	Log message
input.logs[].context	string	no	File or module name
input.strategy	string	no	"auto", "balanced", "innovate", "harden", "repair-only"
input.target_file	string	no	Focus analysis on a specific file

Output (Analyze)

Field	Type	Description
patterns	array	Detected error/regression/inefficiency patterns with severity
health_score	number	System health 0�?00
recommendations	string[]	Actionable improvement suggestions
summary	object	Counts: total_logs, error_count, warn_count, unique_patterns

Output (Evolve)

Field	Type	Description
evolution_id	string	Unique proposal ID
strategy	string	Effective strategy used
recommendations	array	Prioritized improvements with category and approach
risk_assessment	object	Risk level and contributing factors
estimated_improvement	string	Projected health score improvement

Error Codes

• 400 — Invalid action or missing logs array
• 500 — Processing failed

About Claw0x

Claw0x is the native skills layer for AI agents — providing unified API access, atomic billing, and quality control.

Explore more skills: claw0x.com/skills

GitHub: github.com/kennyzir/capability-evolver

---

Deterministic vs LLM Analysis: Which is Right for You?

Feature	LLM-Based (GPT-4, Claude)	Capability Evolver (Local)
Setup Time	5-10 min (prompt engineering)	30 seconds (install skill)
Processing Speed	5-30 seconds	Sub-100ms
Reproducibility	❌ Varies per run	✅ Same logs = same results
Hallucination Risk	⚠️ Can invent patterns	✅ Only reports real patterns
Cost	$0.10-0.50 per analysis	Free (runs locally)
Semantic Understanding	✅ Understands context	❌ Pattern-based only
Audit Trail	❌ Hard to explain	✅ Rule-based, explainable
Privacy	⚠️ Sends data to API	✅ Runs entirely locally

When to Use LLM-Based

• Need semantic understanding of log messages
• Want natural language explanations
• Logs contain unstructured text
• Willing to trade speed for insight depth

When to Use Capability Evolver (Local)

• Need reproducible results for compliance
• Want sub-second processing
• Building automated pipelines
• Require explainable AI for audits
• Processing millions of logs
• Privacy-sensitive applications
• Zero-cost operations

---

How It Fits Into Your Agent Lifecycle

┌─────────────────────────────────────────────────────────────┐
│                  Agent Development Lifecycle                 │
└─────────────────────────────────────────────────────────────┘
                            │
                            ├─ Development
                            │  • Write agent code
                            │  • Local testing
                            │
                            ├─ Staging Deployment
                            │  agent.run('capability-evolver', 
                            │    {action: "analyze", logs: staging_logs})
                            │  → Health check before production
                            │
                            ├─ Production Monitoring
                            │  agent.run('capability-evolver', 
                            │    {action: "analyze", logs: recent_logs})
                            │  → Real-time health tracking (every hour)
                            │
                            ├─ Incident Response
                            │  agent.run('capability-evolver',
                            │    {action: "analyze", logs: incident_logs})
                            │  → Root cause analysis
                            │
                            └─ Continuous Improvement
                               agent.run('capability-evolver',
                                 {action: "evolve", strategy: "balanced"})
                               → Auto-generate improvement tasks (daily)

Integration Points

1. Pre-Deployment — Health check before releasing
2. Real-Time Monitoring — Continuous health tracking
3. Incident Response — Fast root cause analysis
4. Daily Reviews — Automated improvement proposals
5. Fleet Management — Cross-agent pattern detection

---

Why Use Capability Evolver?

Zero-Cost Operations

• Runs locally — no API calls, no billing
• Complete privacy — logs never leave your system
• Offline capable — works without internet connection

Agent-Optimized

• Deterministic analysis — reproducible, auditable results
• Fast processing — sub-100ms, suitable for real-time monitoring
• Structured output — JSON format, easy to integrate
• Evolution strategies — tailored recommendations based on context

Production-Ready

• No dependencies — pure logic, no external services
• Scales to millions — handle enterprise-scale log analysis
• Cloud-native — works in Lambda, Cloud Run, containers
• Zero maintenance — no model updates or API keys to manage