Create or edit a high-level plan for a complex feature. Use when the design is clear but needs documentation, or to convert a brainstorm into an actionable plan.
From cortexnpx claudepluginhub jsvitolo/cortex-plugins --plugin cortexThis skill is limited to using the following tools:
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.
Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.
This skill creates or edits a high-level plan for documenting approach and design before implementation.
$ARGUMENTS
Can be:
/plan "Refactor payment module"/plan PL-5/plan BS-3Convert the brainstorm to a plan:
mcp__cortex__brainstorm(action="to_plan", id="BS-N")
This creates a new plan with the brainstorm's selected ideas and decisions.
Get the existing plan:
mcp__cortex__highlevel_plan(action="get", id="PL-N")
Create a new plan:
mcp__cortex__highlevel_plan(action="create", title="$ARGUMENTS")
A good plan includes:
## Overview
Brief description of what this plan achieves.
## Goals
- Goal 1
- Goal 2
## Approach
How we will implement this:
1. Step one
2. Step two
3. Step three
## Technical Design
### Architecture
- Component A: responsibility
- Component B: responsibility
### Data Model
- Entity changes
- New tables/fields
### API Changes
- New endpoints
- Modified endpoints
## Tasks
- [ ] Task 1: description
- [ ] Task 2: description
- [ ] Task 3: description
## Risks & Mitigations
| Risk | Mitigation |
|------|------------|
| Risk 1 | How to handle |
## Open Questions
- Question 1?
- Question 2?
Update the plan content:
mcp__cortex__highlevel_plan(action="update", id="PL-N", content="<markdown content>")
If this plan relates to an epic:
mcp__cortex__highlevel_plan(action="link_epic", id="PL-N", epic_id="CX-E")
When the plan is ready:
mcp__cortex__highlevel_plan(action="update", id="PL-N", status="approved")
User: /plan Implement caching layer
Claude: Creates PL-1 "Implement caching layer"
Asks: "What's the main goal? Performance? Reducing DB load?"
User: Both, we're getting too many DB hits on the product catalog
Claude: Writes plan with:
- Overview: Add Redis caching for product catalog
- Goals: Reduce DB load by 80%, sub-100ms response times
- Approach: Cache-aside pattern with TTL
- Technical Design: Redis, cache keys, invalidation strategy
- Tasks: Setup Redis, Add cache layer, Add invalidation, Add metrics
User: Looks good, but add a task for testing cache invalidation
Claude: Updates plan with new task
Asks: "Ready to approve and create tasks?"
User: Yes, approve it
Claude: Approves PL-1
Offers to create tasks from the plan
This skill is interactive - engage with the user to:
After planning, the user can:
/implement CX-N - Start implementing