skill-writer_skill

Overview

This skill generates complete, valid Warden skill YAML definitions from natural-language descriptions. It turns a described purpose into a ready-to-run skill file that includes prompt text, allowed/denied tools, and SkillReport expectations. The output follows the required schema and severity guidelines so the skill can be used in automation or CI.

How this skill works

The skill reads the user’s natural-language description and clarifies purpose, scope, and targets. It composes a precise system prompt for the analysis agent, selects appropriate allowed and denied tools, and encodes severity rules and the SkillReport output schema. Finally it emits a linted YAML skill definition and a short rationale for design choices.

When to use it

• You want to create a new Warden skill from a plain-language requirement.
• Standardizing review checks across teams or projects.
• Automating security, style, or test checks in CI pipelines.
• Rapidly prototyping rule-based code analysis agents.
• Converting compliance requirements or playbooks into executable checks.

Best practices

• Start with a concise purpose statement describing what to analyze and why.
• Specify which files, directories, or patterns the skill should inspect (use Glob).
• List allowed and denied tools explicitly to limit agent capabilities.
• Map findings to the provided severity levels and supply clear criteria for each level.
• Include example findings and at least one SkillReport output snippet for validation.

Example use cases

• Create a security-audit skill that checks dependencies and insecure patterns and uses WebFetch for CVE lookups.
• Generate a test-coverage skill that inspects new code and flags untested public APIs.
• Produce a linting skill that searches for banned APIs and enforces project conventions.
• Build a CI check that reports SkillReport JSON for gating pull requests.
• Convert a compliance checklist into a reproducible skill with explicit severity rules.