qe-chaos-resilience — Claude Code skill: chaos & resilience testing
qe-chaos-resilience — a skill for Claude Code, Codex, Cursor and other agents that injects controlled faults into distributed systems and checks recovery behavior. Runs during circuit breaker, failover, and retry logic validation.
MIT 123 installs ★ 533 updated 2026-10-25
npx skills add proffesor-for-testing/agentic-qe --skill qe-chaos-resilience -a claude-codeWhat qe-chaos-resilience does
The skill equips the agent with aqe chaos run, aqe chaos load, aqe chaos stress, and aqe chaos circuit-breaker commands, delegating execution to dedicated sub-agents: qe-chaos-engineer, qe-load-tester, and qe-resilience-tester.
Six fault types are covered: network latency, packet loss, CPU stress, memory pressure, disk fill, and process kill. Load scenarios follow a ramp-up / steady-state / ramp-down profile, and stress tests step through user counts to locate the breaking point. After each experiment the agent produces a report with before/during/after metric snapshots, recovery time, and severity-rated findings.
Safety controls cap blast radius at 10% of affected pods and trigger automatic rollback when error rate exceeds 50% or p99 latency goes above 10 seconds. Environments tagged production-critical are excluded, and production runs require two approvals — so the skill will not fire against live critical systems without explicit sign-off.
What it helps with
- verify that a circuit breaker trips under 500ms network latency
- run a peak-traffic load test against an SLA at 5000 concurrent users
- find the breaking point of
/api/checkoutwith step-increase stress testing - confirm automatic recovery from pod termination within 30 seconds
- validate data integrity across a
database-failoverscenario
Example prompts for the agent
- Run a chaos experiment on
api-service: inject 500ms latency and report recovery time - Execute the
peak-trafficload test, ramp to 10,000 users over 5 minutes, hold for 30 minutes, compare results against SLAs - Test the circuit breaker for
payment-serviceand generate a full chaos report with severity findings
How to install qe-chaos-resilience
1Skills CLI (Vercel)
npx skills add proffesor-for-testing/agentic-qe --skill qe-chaos-resilience -a claude-code2Manually: put the files in place
~/.claude/skills/qe-chaos-resilience/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/qe-chaos-resilience/
1Skills CLI (Vercel)
npx skills add proffesor-for-testing/agentic-qe --skill qe-chaos-resilience -a codex2Manually: put the files in place
.agents/skills/qe-chaos-resilience/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add proffesor-for-testing/agentic-qe --skill qe-chaos-resilience -a cursor2Manually: put the files in place
.cursor/skills/qe-chaos-resilience/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install proffesor-for-testing/agentic-qe assets/skills/qe-chaos-resilience2Manually: put the files in place
.github/skills/qe-chaos-resilience/SKILL.md1Skills CLI (Vercel)
npx skills add proffesor-for-testing/agentic-qe --skill qe-chaos-resilience -a gemini-cli2Platform CLI
gemini skills install https://github.com/proffesor-for-testing/agentic-qe --path assets/skills/qe-chaos-resilienceManually: put the files in place
.devin/skills/qe-chaos-resilience/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add proffesor-for-testing/agentic-qe --skill qe-chaos-resilience -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about qe-chaos-resilience
How do I install qe-chaos-resilience in Claude Code?
Run npx skills add proffesor-for-testing/agentic-qe --skill qe-chaos-resilience -a claude-code. The skill file will be placed at ~/.claude/skills/qe-chaos-resilience/SKILL.md.
What does qe-chaos-resilience do?
It teaches the agent to run fault injection experiments, load and stress tests, and recovery validation for distributed systems. Each experiment produces a structured report with metric snapshots, recovery time, and severity-rated recommendations.
Does the skill protect production environments?
Environments marked production-critical are excluded from all experiments. Any other production target requires two approvals before the agent proceeds, and automatic rollback fires when error rate tops 50% or p99 latency exceeds 10 seconds.
Source text
SKILL.md · MIT
---
name: "qe-chaos-resilience"
description: "Injects controlled faults (network partition, latency, process kill, disk pressure) into distributed systems and validates recovery behavior. Use when testing circuit breakers, failover paths, retry logic, or building confidence in system resilience through chaos engineering."
trust_tier: 3
validation:
schema_path: schemas/output.json
validator_path: scripts/validate-config.json
eval_path: evals/qe-chaos-resilience.yaml
---
# QE Chaos Resilience
## Purpose
Guide the use of v3's chaos engineering capabilities including controlled fault injection, load/stress testing, resilience validation, and disaster recovery testing.
## Activation
- When testing system resilience
- When performing chaos experiments
- When load/stress testing
- When validating disaster recovery
- When testing circuit breakers
## Quick Start
```bash
# Run chaos experiment
aqe chaos run --experiment network-latency --target api-service
# Load test
aqe chaos load --scenario peak-traffic --duration 30m
# Stress test to breaking point
aqe chaos stress --endpoint /api/users --max-users 10000
# Test circuit breaker
aqe chaos circuit-breaker --service payment-service
```
## Agent Workflow
```typescript
// Chaos experiment
Task("Run chaos experiment", `
Execute controlled chaos on api-service:
- Inject 500ms network latency
- Monitor service health metrics
- Verify circuit breaker activation
- Measure recovery time
- Document findings
`, "qe-chaos-engineer")
// Load testing
Task("Performance load test", `
Run load test simulating Black Friday traffic:
- Ramp up to 10,000 concurrent users
- Maintain load for 30 minutes
- Monitor response times and error rates
- Identify bottlenecks
- Compare against SLAs
`, "qe-load-tester")
```
## Chaos Experiments
### 1. Fault Injection
```typescript
await chaosEngineer.injectFault({
target: 'api-service',
fault: {
type: 'latency',
parameters: {
delay: '500ms',
jitter: '100ms',
percentage: 50
}
},
duration: '5m',
monitoring: {
metrics: ['response_time', 'error_rate', 'throughput'],
alerts: true
},
rollback: {
automatic: true,
trigger: 'error_rate > 10%'
}
});
```
### 2. Load Testing
```typescript
await loadTester.execute({
scenario: 'peak-traffic',
profile: {
rampUp: '5m',
steadyState: '30m',
rampDown: '5m'
},
users: {
initial: 100,
target: 5000,
pattern: 'linear'
},
assertions: {
p95_latency: '<500ms',
error_rate: '<1%',
throughput: '>1000rps'
}
});
```
### 3. Stress Testing
```typescript
await loadTester.stressTest({
endpoint: '/api/checkout',
strategy: 'step-increase',
steps: [100, 500, 1000, 2000, 5000],
stepDuration: '5m',
findBreakingPoint: true,
monitoring: {
resourceUtilization: true,
databaseConnections: true,
memoryUsage: true
}
});
```
### 4. Resilience Validation
```typescript
await resilienceTester.validate({
scenarios: [
'database-failover',
'cache-failure',
'external-service-timeout',
'pod-termination'
],
expectations: {
gracefulDegradation: true,
automaticRecovery: true,
dataIntegrity: true,
recoveryTime: '<30s'
}
});
```
## Fault Types
| Fault | Description | Use Case |
|-------|-------------|----------|
| Latency | Add network delay | Test timeouts |
| Packet Loss | Drop network packets | Test retry logic |
| CPU Stress | Consume CPU | Test resource limits |
| Memory Pressure | Consume memory | Test OOM handling |
| Disk Full | Fill disk space | Test disk errors |
| Process Kill | Terminate process | Test recovery |
## Chaos Report
```typescript
interface ChaosReport {
experiment: {
name: string;
target: string;
fault: FaultConfig;
duration: number;
};
results: {
hypothesis: string;
validated: boolean;
metrics: {
before: MetricSnapshot;
during: MetricSnapshot;
after: MetricSnapshot;
};
events: ChaosEvent[];
recovery: {
detected: boolean;
time: number;
automatic: boolean;
};
};
findings: {
severity: 'critical' | 'high' | 'medium' | 'low';
description: string;
recommendation: string;
}[];
artifacts: {
logs: string;
metrics: string;
traces: string;
};
}
```
## Safety Controls
```yaml
safety:
blast_radius:
max_affected_pods: 1
max_affected_percentage: 10
abort_conditions:
- error_rate > 50%
- p99_latency > 10s
- service_unavailable
excluded_environments:
- production-critical
required_approvals:
production: 2
staging: 0
```
## SLA Validation
```typescript
await resilienceTester.validateSLA({
slas: {
availability: 99.9,
p95_latency: 500,
error_rate: 0.1
},
period: '30d',
report: {
breaches: true,
trends: true,
projections: true
}
});
```
## Coordination
**Primary Agents**: qe-chaos-engineer, qe-load-tester, qe-resilience-tester
**Coordinator**: qe-chaos-coordinator
**Related Skills**: qe-performance, security-testing
© proffesor-for-testing · MIT · proffesor-for-testing/agentic-qe