Maps of Agent Skill Sets: What Each Does and When to Use It
Each article is an interactive map of one coding agent skill set: what it is for, how it differs from the others, and when to use it.
4 articles
Pick by what you need
How do you want to work with the agent?
Four skill sets, side by side
Compare:
| Criterion | Agent Skills | Superpowers | pstack | mattpocock/skills |
|---|---|---|---|---|
| Shape | Six project stages | One mandatory sequence of steps | Playbooks chosen by task type | Small skills you combine |
| How it starts | A command per stage, skills auto-activate | Starts automatically | One command that calls the rest | You type commands, some skills start automatically |
| Strongest at | Skills for the whole lifecycle | Long, reviewed autonomous runs | Evidence and cross-model checks | Requirements and control |
| The cost | You move through each stage yourself | Too much process for small changes | Uses many tokens, some parts only run on Cursor | You assemble the process yourself |
In this series
A map of each skill set
Agent Skills by Addy Osmani: 25 Skills Across 6 Stages, Mapped
Addy Osmani's Agent Skills on one map: 25 skills across six project stages, how it differs from other sets, and when to use it.
Superpowers, Mapped: The 7-Step Process That Disciplines Agents
Superpowers on one map: the seven steps an agent must follow, which ones block it, how it differs from other skill sets, and when to use it.
mattpocock/skills, Mapped: Four Agent Failures and Their Fixes
mattpocock/skills on one map: start from what your agent gets wrong, find the matching skill, and see how the set differs from the others.
pstack, Mapped: One Command That Picks the Right Playbook
pstack by poteto on one map: pick a type of work, see the steps /poteto-mode runs, how it differs from other skill sets, and when to use it.