synthetic-monitoring-api

name	synthetic-monitoring-api
description	Set up synthetic monitoring for an API. Use this skill when a user wants to create automated end-to-end tests that run on a schedule to continuously verify API health and discover issues before end users do.
argument-hint	<api-docs-url> [<guide-url>]
model	sonnet

Synthetic Monitoring for API

Critical rules

• This is a complex, multi-step skill that requires careful task management. Whenever you encounter a to-do item in this file or any reference guide, immediately add it to your task list.
• Guides in references/ are binding rules, not optional guidance. You must follow them precisely.
• Prefer subagents for consuming large content (crawl output, user guides, PDFs) or authoring code — keeps main context clean. Carry forward all relevant constraints into subagent prompts; never assume the subagent knows rules from references you read.

Workflow:

1. Gather API context from user-provided sources
2. Generate API spec
3. Generate testing spec
4. Collect necessary information for coding and deployment
5. Implement the code
6. Deploy
7. Conclude

Details for each step

Step 1: Gather context

$ARGUMENTS may contain up to two positional URLs: API documentation (first), user/help guide (second). The guide URL adds end-user context that improves test realism — explain this when asking for it.

Collect information about the user's API. Support any combination of:

• URLs — crawl via crawl4ai MCP. Read references/web-crawling-guide.md for guidance.
• Files (PDF, markdown, OpenAPI specs) — read tool
• Code — use an explore or similar subagent to explore the codebase and gather API surface

Ask the user for anything missing.

Step 2: API spec

Desired state: specs/api_spec.md and specs/openapi.yaml accurately reflect the full current API surface.

• Read references/api-spec-guide.md and examples at assets/examples/specs/.

Step 3: Testing spec

Desired state: specs/testing_spec.md covers every operation in specs/openapi.yaml.

• Read references/testing-spec-guide.md and the example at assets/examples/specs/testing_spec.md.

Step 4: Collect info for coding and deployment

Desired state: all credentials in specs/.secrets.env, deployment method chosen.

• Read references/collect-code-requirements-from-user.md.

Step 5: Implement code

Desired state: test code matches specs/testing_spec.md exactly.

• Read references/code-authoring-guide.md.

Step 6: Deploy

Read the deployment guide matching the method chosen in Step 4:

• GitHub Actions: references/deployment-github-actions-guide.md
• Docker: references/deployment-docker-guide.md

Step 7: Conclude

Summarize what was accomplished across the previous steps. Tell the user to import https://gist.github.com/font44/b5ad1b520d87ae7356b6583b9b4f6fdc into their Grafana account to see metrics.