smithery/rjroy

prep-plan

Use when ready to build a reviewable implementation plan as a lore artifact.

Installation

$ npx skills add smithery/rjroy --skill prep-plan

Summary

  • Use when ready to build a reviewable implementation plan as a lore artifact.
  • Input can be a spec, design, brainstorm, research, or a well-defined prompt.
  • Triggers include "prep plan", "prepare a plan", "plan this", "make a plan", "plan the implementation", and "break this into steps".

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from smithery/rjroy.

npx skills add smithery/rjroy

Browse all from smithery/rjroy

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,366 B
  • docs SUMMARY.md 240 B

History

  1. First recorded snapshot · 0 installs

SKILL.md

Prep Plan

Build an implementation plan the user can review and approve before any code is written. The plan is the output. Implementation is a separate step via /implement.

Do not use EnterPlanMode. This skill authors a document, not a precursor to immediate code changes.

Process

Gather context. Invoke the lore-researcher agent via Task tool with the topic description — wait for the result. Read any relevant lore artifacts (spec, design, brainstorm, research). Explore the codebase with an Explore subagent to understand what exists and where changes will land.

Surface gaps before drafting. If the input has ambiguous requirements, contradictions, or unstated assumptions that would force guesses, list them and ask the user to resolve them first. A plan built on unreviewed assumptions fails during implementation.

Draft collaboratively. Map the goal or requirements to concrete implementation steps. Order by dependency. Flag steps that need specialized expertise. Include a final validation step that checks the implementation against the source artifact.

Save. The saved file is the review vehicle — don't present the plan in chat before saving.

Fresh-eyes review. After saving, invoke the plan-reviewer agent via Task tool on the saved plan. Present findings and offer to address issues before implementation begins.

Saving

Save to .lore/work/plans/[feature-name].md using kebab-case. Load ${CLAUDEPLUGINROOT}/shared/frontmatter-schema.md for the frontmatter fields. The document body is freeform — what matters is that steps are concrete (files, functions, changes) and the source artifact or goal is clearly referenced so validation has something to check against.

Write the body Markdown-first per the "Body Format" section of ${CLAUDEPLUGINROOT}/shared/frontmatter-schema.md. Reach for embedded inline HTML when the step flow needs it — a visual step sequence with dependency indicators, and validation gates distinguished visually from the step description.