Home/Skills/meta/Writing Plans

/writing-plans

Writing Plans

Turn an approved spec into bite-sized tasks an engineer with zero context could ship. DRY, YAGNI, TDD, frequent commits.

Used by 1 agent

What it does

Implementation-plan authoring for a fresh engineer or a subagent. Maps the files that will be created or modified, decomposes into tasks at 2-to-5 minute granularity, writes each task as a TDD cycle (failing test, verify red, minimal code, verify green, refactor, commit), includes exact file paths and test names, and hands off cleanly to execution.

Use it after `brainstorming` approves a spec and before any code gets written. Especially valuable when the execution session will be a different agent or a different day, because the plan has to carry the full context alone.

Outputs a plan at `docs/superpowers/plans/YYYY-MM-DD-feature-name.md` with the required header, file structure, and task-by-task checkboxes. Pairs with `brainstorming` (the upstream approval gate), `executing-plans` (sequential execution), and `subagent-driven-development` (parallel execution with per-task subagents).

Build your second brain

Run /writing-plans with your agents.

This skill ships with Agent0 Starter, plus 82 others. Cortex runs it from day one.

Cortex
Chief of Staff
Loom
Loom
Radar
Radar
Hippo
Hippocampus
Signal
Signal
Sentinel
Sentinel
Axon
Axon
Hey, I'm Cortex — Chief of Staff for the team. Ask me anything; I'll handle it or pull in Loom (writing), Radar (growth), Hippocampus (research), Signal (social), Sentinel (ops), or Axon (code review). What's on your mind?