Skill Smokeplanner Skill
When To Use
Use this skill when reviewing, packaging, or releasing an agent skill and you need a deterministic local smoke plan from its SKILL.md.
Required Tools Or Inputs
- Node.js 20 or newer.
- A local
SKILL.mdfile. - Optional nearby
package.jsonwith verification scripts.
Side-Effect Boundaries
The planner reads files and prints Markdown or JSON. It does not run shell commands, install packages, publish packages, send messages, or call external APIs.
Approval Requirements
Ask for approval before running any generated command that may mutate a Git repository or GitHub resource, publish or version a package, deploy, send messages, remove system paths, or call a live service.
Examples
Shell examples may use backtick or tilde fences of three or more matching markers. The closing fence must use the same marker and be at least as long as the opening fence.
skill-smokeplanner plan ./SKILL.md
skill-smokeplanner plan ./SKILL.md --json
skill-smokeplanner plan ./skills/example/SKILL.md --repo-root .
Without --repo-root, package metadata is discovered upward from SKILL.md
and is bounded to its containing Git worktree. Pass --repo-root when the
intended package root is elsewhere or an explicit boundary is preferable.
Required H2 sections may organize their content under H3-H6 subsections. The planner keeps each child subsection in the parsed output and also attributes its visible content to the enclosing H2. Fenced heading examples remain ignored, and a new H2 starts a separate section boundary.
Validation Workflow
Run npm test, npm run check, and npm run smoke. Confirm taxonomy fixtures warn for repository, GitHub CLI, package publish/version, and network commands in direct examples and preferred package-script bodies; confirm safe-neighbor fixtures do not trigger broad substring matches.