Your AI API tester that validates endpoints, monitors health, and tracks performance.
Probe keeps your API reliable:
- Tests endpoints with various inputs and edge cases
- Runs periodic health checks
- Tracks response time trends
- Alerts on failures and slow endpoints
mkdir -p ~/.openclaw/agents/api-tester/agent
cp SOUL.md ~/.openclaw/agents/api-tester/agent/
openclaw agents add api-tester --workspace ~/.openclaw/agents/api-testeropenclaw chat api-tester "Health check all endpoints"You: "Test the checkout API"
Probe: [4 tests with inputs, responses, pass/warn/fail, suggestions]
You: "Health check all endpoints"
Probe: [Table of endpoints with status, response time, result]
You: "Generate tests for this OpenAPI spec"
Probe: [Happy path, validation, edge case, and security tests]
You: "Which endpoints are slow?"
Probe: [Endpoints exceeding threshold with trend data]
POST /api/stripe/checkout
Test 1: Valid request → 200 (340ms) PASS
Test 2: Missing type → 200 (125ms) WARN
Test 3: Invalid type → 500 (89ms) WARN
Test 4: Empty body → 500 (45ms) WARN
Summary: 1 pass, 3 warnings
Suggestion: Add input validation
API Health - Feb 16
GET / 200 45ms PASS
POST /checkout 200 340ms PASS
POST /generate 200 890ms SLOW
GET /health 200 22ms PASS
5/5 up | Avg 283ms
Alert: /generate > 500ms threshold
- Test after every deploy - Catch issues before users do
- Set response time thresholds - Know when things slow down
- Include edge cases - Empty inputs, large payloads, special characters
- Monitor trends - A slow endpoint today is a down endpoint tomorrow
- v1.0.0 - Initial release with endpoint testing
- v1.1.0 - Health monitoring and alerts
- v1.2.0 - OpenAPI test generation
Created by @openclaw
MIT