plan

Writes actionable markdown implementation plans with bite-sized tasks to .nyxora/plans/ without executing code.

5|2|Updated May 26, 2026
One-click install
npx skills add https://github.com/perasyudha/Nyxora --skill plan-perasyudha
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: plan
Source: https://github.com/perasyudha/Nyxora/tree/main/packages/core/playbooks/software-development/plan
Command: npx skills add https://github.com/perasyudha/Nyxora --skill plan-perasyudha

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Jumping straight into coding complex features without a structured plan leads to missed requirements, vague tasks, and rework. This Skill enforces a plan-only mode that produces a concrete, step-by-step implementation plan before any code is written. ## Core Features & Use Cases - Plan-Only Mode: Restricts the agent to read-only inspection and writing a single markdown plan file, with no code edits, commits, or mutating commands. - Structured Plan Format: Generates plans with exact file paths, complete copy-pasteable code, TDD steps, verification commands, and bite-sized tasks of 2-5 minutes each. - Timestamped Storage: Saves plans under .nyxora/plans/ with timestamped filenames, keeping them attached to the active workspace across local, docker, ssh, and other backends. - Use Case: Before implementing a multi-step feature like user authentication, invoke plan mode to produce a task-by-task breakdown with failing tests, minimal implementations, and commit steps that a subagent can execute later. ## Quick Start Ask the agent to enter plan mode and write an implementation plan for your feature, for example: plan out adding password reset functionality to the auth module.

Frequently Asked Questions about plan

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a good software implementation plan before coding?

Break the feature into bite-sized tasks of 2-5 minutes each, with exact file paths, complete code examples, and verification commands. Follow a TDD cycle per task: write a failing test, verify it fails, implement minimal code, verify it passes, then commit.

How to make an AI agent plan instead of writing code?

Invoke plan mode, which restricts the agent to read-only inspection and writing a single markdown plan file. The agent cannot edit project files, run mutating commands, or commit until the plan is approved and execution begins.

Where are the generated plan files saved?

Plans are saved under .nyxora/plans/ with a timestamped filename like YYYY-MM-DD_HHMMSS-slug.md. The path is relative to the active workspace, so it works across local, docker, ssh, modal, and daytona backends.

What should an implementation plan include for subagent execution?

Each task needs an objective, exact files to create or modify, a failing test, the minimal implementation, run commands with expected output, and a commit step. Plans should reference subagent-driven-development for task-by-task execution with spec and quality reviews.

When should I skip writing a full implementation plan?

The skill recommends almost never skipping it, since even simple features hide assumptions that cause bugs. Plans are especially important before multi-step features, complex requirements, or delegating work to subagents.