Claude Skills
CollectionsCompareWorkflowsNominate
Sign inSign up
© 2026 Curated Agent Skills·Learn more about Agent Skills
All curated skills
Expert collection · 14 stable skills

Jesse Vincent Superpowers

A complete agent development methodology that plans before it codes and verifies before it claims success.

View original repository

How this collection thinks

Superpowers is a methodology rather than a menu. Its skills are meant to fire automatically at the moments that decide whether agent work succeeds: brainstorming and design sign-off before implementation, a written plan detailed enough for a context-free engineer, strict red-green test-driven development, subagent execution with per-task review, and verification evidence before anyone claims success. The through-line is process over improvisation and evidence over assertion, enforced by a rule that the agent must check for an applicable skill before it responds.

Install

/plugin install superpowers@claude-plugins-official

Best for

Developers and teams who want an agent that plans before coding, keeps changes small and reviewed, and refuses to call work done without fresh evidence, especially in Claude Code, Codex, Cursor, and other harnesses with plugin or skill support.

Trade-offs

The methodology is deliberately heavy. Brainstorming, chunked design review, and a full written plan add real friction to small or already-clear changes, and the automatic skill-invocation rule can feel rigid. Its richest behavior assumes a harness with subagents, hooks, and a plugin marketplace, so the experience is uneven across agents even though the skills themselves travel well.

Start here

Our reviewed entry points

These are the skills with the clearest standalone value. Each review explains the benefit and the cost.

Jesse Vincent
Design

Brainstorming

The entry skill is the reason the rest of the methodology works. Brainstorming refuses to let the agent jump into code, classifies how much process a request needs, and drives the idea through context, refinement, a chunked design, and explicit human sign-off before anything is planned or built.

SourceRead review
Jesse Vincent
Debugging

Systematic Debugging

A four-phase root-cause process that treats symptom fixes as failures. It forces the agent to reproduce, isolate, understand, and only then change code, which is exactly the discipline most agents skip when they start guessing at a fix.

SourceRead review
Jesse Vincent
Verification

Verification Before Completion

The counterweight to optimistic agents: it requires fresh command output as evidence before any claim that work is fixed, passing, or complete. Pairing it with a planning or implementation skill removes most false success reports.

SourceRead review

Stable inventory

All 14 curated skills

In-progress experiments are excluded from this editorial collection.

Design

Brainstorming

Turns a rough idea into an approved design through Socratic questions, chunked review, and explicit sign-off before any code.

View skill source

Planning

Writing Plans

Writes an implementation plan detailed enough for a junior engineer with no project context to follow task by task.

View skill source

Executing Plans

Loads a written plan, reviews it critically, executes every task in order, and reports at review checkpoints.

View skill source

Testing

Test-Driven Development

Enforces the red-green-refactor cycle and refuses to write implementation code before a failing test is observed.

View skill source

Debugging

Systematic Debugging

Runs a four-phase root-cause process and treats symptom fixes as failures.

View skill source

Verification

Verification Before Completion

Requires fresh evidence from real commands before claiming that work is fixed, passing, or complete.

View skill source

Orchestration

Subagent-Driven Development

Dispatches a fresh implementer per task with spec and quality review, then finishes with a whole-branch review.

View skill source

Dispatching Parallel Agents

Splits independent work across isolated subagents so each receives crafted context instead of the parent session history.

View skill source

Review

Requesting Code Review

Sends a change to a context-isolated reviewer before it can cause cascading problems.

View skill source

Receiving Code Review

Evaluates review feedback with technical rigor instead of performative agreement or blind implementation.

View skill source

Git

Using Git Worktrees

Creates an isolated workspace with native tooling or a git worktree before feature work begins.

View skill source

Finishing a Development Branch

Verifies tests, detects the environment, and presents a merge, pull request, or cleanup decision for completed work.

View skill source

Agent Design

Writing Skills

Treats skill authoring as test-driven development for process documentation, including how to verify a skill actually works.

View skill source

Workflow

Using Superpowers

Establishes the mandatory habit of checking for an applicable skill before any response and explains the system.

View skill source

Not sure which collection fits?

Explore the other expert toolkits by workflow, scope, platform, and strongest use cases.

Explore all collections