How do I install the Plan skill?

Run npx skillmds@latest add lucassantana-dev/plan in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Plan skill do?

Build a compact, validation-gated implementation plan for multi-step, risky, or ambiguous work — typically when phases need explicit checkpoints, the change touches multiple files/services, or the user wants alignment before execution. Output goes to `.claude/plans/ .md` (or `.agents/plans/`) with goal, in/out-of-scope, phased steps, validation per phase, and replanning triggers. Skip for trivial single-file fixes, bug investigations (use `tracer`), or open-ended exploration (use `explore`). Pair with `ship` once the plan ends and merge work begins. It is listed under Coding & Dev Tools on SkillMD.

SkillMD's automated safety review verdict for this skill is PASS. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: docs only, makes network calls. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Plan?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

LucasSantana-Dev (@lucassantana-dev) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Plan by lucassantana-dev — Agent Skill

Name: Plan
Rating: 4.2 (5 reviews)
Author: LucasSantana-Dev

Plan

Build a compact, validation-gated implementation plan for multi-step, risky, or ambiguous work — typically when phases need explicit checkpoints, the change touches multiple files/services, or the user wants alignment before execution. Output goes to `.claude/plans/<topic>.md` (or `.agents/plans/`) with goal, in/out-of-scope, phased steps, validation per phase, and replanning triggers. Skip for trivial single-file fixes, bug investigations (use `tracer`), or open-ended exploration (use `explore`). Pair with `ship` once the plan ends and merge work begins.

LucasSantana-Dev Updated Aug 19, 2026 1 repo stars

File contents

plan

Use planning only when it actually reduces risk.

Steps

Read local guidance and active handoff.
Pull only the context needed for the task.
State goal, in-scope, and out-of-scope.
Break work into short phases.
Give each phase a validation step.
Write the plan to .claude/plans/ or .agents/plans/.

Rules

Keep phases small enough to finish without drifting.
Note what would cause replanning.
If work is already partly done, document the current state before planning the rest.
Prefer the provided plan template.

Extend vs new plan

Before writing, scan .claude/plans/ and .agents/plans/:

Existing plan covers same scope AND ≤30 days old → extend (append phases, edit status header, keep history).
Same scope but >30 days old → new plan; mark old as superseded.
Different scope → new plan.

When extending, update the status header to reflect partial completion (e.g. "Phases 1-5 done; Phases 6-8 added 2026-05-14").

Skip plan if

≤2 files touched AND edit is obvious from the request → just do it, no plan
Pure investigation / "where is X" → use explore
Bug root-cause hunt → use tracer
Open-ended ideation → use brainstorming

Worked example — extend

User: "add WoL panel + tabs to dashboard." Existing: homepage-customization-2026-05-13.md (5 phases, 1 day old). Action: append Phase 6-8 to that file; do NOT create a second plan. Update header to note the extension date.

---name: plandescription: Build a compact, validation-gated implementation plan for multi-step, risky, or ambiguous work — typically when phases need explicit checkpoints, the change touches multiple files/services, or the user wants alignment before execution. Output goes to `.claude/plans/<topic>.md` (or `.agents/plans/`) with goal, in/out-of-scope, phased steps, validation per phase, and replanning triggers. Skip for trivial single-file fixes, bug investigations (use `tracer`), or open-ended exploration (use `explore`). Pair with `ship` once the plan ends and merge work begins.---# planUse planning only when it actually reduces risk.## Steps1. Read local guidance and active handoff.2. Pull only the context needed for the task.3. State goal, in-scope, and out-of-scope.4. Break work into short phases.5. Give each phase a validation step.6. Write the plan to `.claude/plans/` or `.agents/plans/`.## Rules- Keep phases small enough to finish without drifting.- Note what would cause replanning.- If work is already partly done, document the current state before planning the rest.- Prefer the provided plan template.## Extend vs new planBefore writing, scan `.claude/plans/` and `.agents/plans/`:- Existing plan covers same scope AND ≤30 days old → **extend** (append phases, edit status header, keep history).- Same scope but >30 days old → new plan; mark old as **superseded**.- Different scope → new plan.When extending, update the status header to reflect partial completion (e.g. "Phases 1-5 done; Phases 6-8 added 2026-05-14").## Skip plan if- ≤2 files touched AND edit is obvious from the request → just do it, no plan- Pure investigation / "where is X" → use `explore`- Bug root-cause hunt → use `tracer`- Open-ended ideation → use `brainstorming`## Worked example — extendUser: "add WoL panel + tabs to dashboard."Existing: `homepage-customization-2026-05-13.md` (5 phases, 1 day old).Action: append Phase 6-8 to that file; do NOT create a second plan. Update header to note the extension date.

Frequently asked questions