Tooling first automation
Skill selamy-labs/agent-skills/skills/tooling-first-automation
Use when a task can be performed through existing callable tooling, MCP servers, browser automation, CLI wrappers, or service APIs before writing bespoke scripts or ad hoc automation. Especially relevant for browser/UI automation, external-system actions, and repeated operational workflows.From its SKILL.md
npx -y skills add selamy-labs/agent-skills --skill tooling-first-automationAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
2.0 KB, 358 tokens by cl100k_base, as published. Nobody here has run it
Tooling-First Automation
Before writing new automation, inspect the tools already available in the environment and prefer the narrowest existing callable path that fits.
Workflow
- Check the configured tool surface: MCP servers, built-in browser tools, repository scripts, CLIs, SDK helpers, and service-specific wrappers.
- Use a first-party or official integration when it covers the task. For browser work, prefer a configured browser automation MCP or built-in browser tool before hand-writing Playwright, Puppeteer, Selenium, or CDP scripts.
- If no existing tool fits, write the smallest scoped script needed and explain why the existing tools were insufficient.
- If the right tool exists but is missing dependencies, broken, or unavailable in the current environment, fix or codify the tool installation/configuration in the owning source of truth before treating a one-off script as normal.
- Verify the real tool path works with a representative action, not just that a package or config entry exists.
When Not To Use
- The user explicitly asks for a from-scratch implementation.
- The existing tool cannot perform the required action safely or with enough observability.
- A short local script is the repository's established interface for the task.
Anti-Patterns
- Writing bespoke browser automation while a browser MCP or built-in browser tool is configured and working.
- Copying an API call into prose or shell snippets when a typed MCP/server tool already exposes the operation.
- Declaring a tool "available" without launching it or completing a real action.
- Leaving a dependency fix as an operator memory instead of codifying it in the source that provisions the environment.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.