Skill writer (Sentry)
Create and improve agent skills.
Meta (skills & agents)
Claude settings audit (Sentry)
Generate recommended Claude Code settings permissions.
Prompt optimizer (Sentry)
Improve prompts using evaluations and context.
Implement a whole spec
Execute an entire spec across multiple concurrent agents.
Skill description tester
Tests whether a skill description will trigger correctly by writing should-trigger and should-not-trigger requests. Use to tune descriptions.
Convert prompt to skill
Turns a prompt you reuse into a proper skill: name, description, instructions and examples. Use to save prompts you repeat.
Agent eval designer
Designs an evaluation for an agent: test tasks, scoring checklist, failure categories and a run plan. Use before shipping an agent.
Mcp server planner
Plans an MCP server: what to expose as tools, inputs and outputs, safety limits and a test plan. Use before building one.
Find skills
Helps your AI find and install the right skill for a task from the open skills ecosystem.
Dispatching parallel agents
Run independent tasks at the same time with separate agents.
Subagent-driven development
Build a plan task by task with fresh subagents and a review between tasks.
Using superpowers
Learn how to find and use the other Superpowers skills.
Writing skills
Create and test new skills.
Skill creator
Create, improve and test new skills.
Write a skill
Guides you through creating a new agent skill: purpose, trigger description, instructions, examples and testing. Use when you want to write a…
Skill reviewer
Reviews a SKILL.md for a clear trigger, focused instructions, safe scripts and testability. Use before publishing a skill.
Agent system prompt writer
Writes a system prompt for an AI agent: role, goals, tools, rules, and when to stop or ask. Use when building an…
Prompt improver
Rewrites a weak prompt into a clear one with context, task, format and examples, and explains the changes. Use when results are…