rune-watchdog

Runs post-deploy HTTP health checks and generates structured smoke test reports.

1|Updated Mar 22, 2026
One-click install
npx skills add https://github.com/dangvu008/VietTruyen --skill rune-watchdog-dangvu008
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: rune-watchdog
Source: https://github.com/dangvu008/VietTruyen/tree/main/.agents/skills/rune-watchdog
Command: npx skills add https://github.com/dangvu008/VietTruyen --skill rune-watchdog-dangvu008

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? After deploying an application, teams need to verify that endpoints are reachable, fast, and error-free before declaring a launch successful. Manual curl checks are inconsistent and produce no structured record of system health. ## Core Features & Use Cases - Automated Health Checks: Runs HTTP status checks against a base URL and endpoint list, classifying each as HEALTHY, REDIRECT, CLIENT_ERROR, SERVER_ERROR, or UNREACHABLE. - Response Time & Performance Analysis: Measures latency per endpoint, flags slow responses, and detects performance regressions by comparing against previous baselines. - Structured Smoke Test Reports: Emits a standardized Watchdog Report with alerts, performance signals, and an overall status of ALL_HEALTHY, DEGRADED, or DOWN. - Use Case: After deploying to production, pass the deployed URL and endpoint list to automatically verify all routes return 2xx within acceptable latency and receive a report flagging any 5xx errors or regressions. ## Quick Start Run a post-deploy health check on https://myapp.com covering the /, /health, and /api/status endpoints and give me the smoke test report.

Frequently Asked Questions about rune-watchdog

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check if my deployed app is healthy after a release?

Provide the deployed base URL and a list of endpoints to check. Each endpoint is probed with curl for HTTP status and response time, then classified as HEALTHY, REDIRECT, CLIENT_ERROR, SERVER_ERROR, or UNREACHABLE in a structured report.

How to detect performance regressions after a deploy?

Response times are compared against the previous watchdog run baseline. Endpoints that were FAST but are now SLOW are flagged as PERF_REGRESSION and correlated with the recent git diff. Without a baseline, the first run is noted as INFO.

What HTTP status codes trigger alerts in smoke tests?

Any 4xx on a non-auth endpoint triggers an alert, any 5xx triggers a critical alert, and an unreachable endpoint (non-zero curl exit) is critical. Response times over 2 seconds are also flagged as alerts.

Can this diagnose why an endpoint is slow?

No, it is a detector, not a diagnoser. It flags performance signals with hypotheses such as N+1 queries or connection pool saturation, but root-cause diagnosis must be handled by a separate performance analysis workflow.

Why is a curl timeout reported as unreachable instead of slow?

A non-zero curl exit code means no response was received, so no valid latency measurement exists. Treating timeouts as slow would corrupt response time data, so they are classified as UNREACHABLE critical alerts instead.