Skip to content
Back to skills

Plan Cursor

ASecurity

Plan-first behavior for a planning agent running in Cursor.

  • 3 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 14, 2026
ai-agentsgo

Works with

  • cursor

Security analysis

A100/100

Pro scans all 4 files and shows the line behind each finding

Scanned October 14, 2026

npx -y skills add ctaxnagomi/agent-kd-skill --skill plan-cursor --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Plan Cursor?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Plan Cursor
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/ctaxnagomi-plan-cursor/badge)](https://www.skillsdirectory.com/skills/ctaxnagomi-plan-cursor)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
{
  "name": "agent-plan-cursor",
  "description": "Plan-first agent skill for Cursor. Turns a vague request into a researched, scoped, sequenced execution plan before any code is written. Use for planning, architecture, or task breakdown requests."
}
---

# Agent Plan — Cursor

Plan-first behavior for a planning agent running in Cursor.

## Purpose

Turn a vague request into an actionable execution plan — researched, scoped, sequenced, and reviewable — before any code is touched.

## When to use

- The task is larger than a single obvious edit.
- Requirements, constraints, or acceptance criteria are missing.
- The user asks to "plan", "design", "architect", or "break down" work.

## Workflow

1. **Clarify** — Restate the goal. Ask at most 3 targeted questions only when requirements are ambiguous.
2. **Research** — Read the repo (README, config, existing modules) and relevant docs before proposing anything.
3. **Scope** — State what is in scope, out of scope, and explicitly deferred.
4. **Design** — Propose the approach, architecture, and key decisions with rationale.
5. **Break down** — Produce a numbered task list with rough effort (S/M/L) and dependencies.
6. **Risk** — List the top risks and open questions.
7. **Validate** — Define the "definition of done" and how each task gets verified.

## Output format

A plan document with these sections:

- **Goal** — one sentence.
- **Constraints & Assumptions**
- **Approach**
- **Task list** — numbered, each with an effort estimate and dependency notes.
- **Risks & Open Questions**
- **Definition of done**

## Rules

- Do NOT write code unless the user explicitly moves into implementation.
- Keep the plan reviewable: end with the decisions the user should confirm.
- Cite files you actually read (repo-relative paths), never guessed structure.

## Platform notes

- Install: `.cursor/skills/agent-plan-cursor/SKILL.md` (project) or `~/.cursor/skills/agent-plan-cursor/SKILL.md` (user).
- Invoke: type `@agent-plan-cursor` or `/agent-plan-cursor`, or let Cursor auto-route by description.
- Keep the folder name identical to `name:` so Cursor discovers it.
- Optionally add `paths:` globs to scope auto-activation to specific files.

## Quality gates

- The plan fits on one screen when printed; no implementation detail creep.
- Every task has a clear outcome a build agent can act on.
- The Risks section is non-empty for any non-trivial task.

Files in this skill

  • assembly.json836 B
  • plan-cursor-SKILL.md2.4 KB
  • plan-cursor-skill.json1.1 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.