ab-test-plan
01Command Claude Code
Design an experiment with the hypothesis, sample size, and decision rule fixed before it runs.
Plug-and-play skills and prompts for every AI coding agent
This repository also configures its own agents. See what AI-SKILLS tells them →
Command Claude Code
Design an experiment with the hypothesis, sample size, and decision rule fixed before it runs.
Command Claude Code
Write advertising copy that leads with the benefit, speaks to one audience, and drives one clear action.
Command Claude Code
Add docstrings and comments that explain why, matching the code's existing style, without noise.
Command Claude Code
Rewrite text in a different tone (more formal, friendlier, more direct) while keeping the meaning.
Command Claude Code
A strict brief that makes a coding agent implement a feature end to end, matching the codebase and verifying it works.
Command Claude Code
A strict, goal-focused brief that makes a coding agent (Claude Code, Cursor, Antigravity) finish a task correctly and verify it.
Command Claude Code
Design an evaluation for an AI agent or LLM feature: what to test, how to grade it, and how to catch regressions.
Command Claude Code
A focused brief that makes a coding agent map an unfamiliar codebase so you (or it) can work in it confidently.
Command Claude Code
A strict brief that makes a coding agent find the root cause of a bug, fix it minimally, and prove it is fixed.
Command Claude Code
A strict brief that makes a coding agent refactor code without changing behavior, verifying nothing broke.
Command Claude Code
Write a system prompt for an AI agent that defines its role, tools, boundaries, and output clearly enough to act reliably.
Command Claude Code
Decompose a complex task into clear subtasks an AI agent (or agents) can execute and verify step by step.
Command Claude Code
Write a tool/function definition for an AI agent with a description that steers use and parameters it fills correctly.
Command Claude Code
Design a multi-agent workflow with the right pattern, explicit handoffs, and human gates on consequential actions.
Command Claude Code
Analyse a competitor's product from public information, separating observation from inference.
Command Claude Code
Diagnose a performance problem from measurements rather than intuition, and propose the fix with the largest expected return.
Command Claude Code
Design a clean REST or GraphQL API from requirements, with sensible resources, methods, status codes, and error shapes.
Command Claude Code
Generate accurate reference documentation for an API from its code, marking every inference.
Command Claude Code
Write a guide that gets consumers from an old API version to a new one with every breaking change addressed.
Command Claude Code
Write a sincere, professional apology email that owns the mistake, addresses the impact, and offers a fix.
Command Claude Code
Write an ADR that captures a technical decision, its context, and its consequences so future teams know why.
Command Claude Code
Assess technical debt by what it costs now, and produce a paydown plan tied to upcoming work.
Command Claude Code
Audit a product against real assistive technology use, prioritised by who is blocked.
Command Claude Code
Run a full autonomous research investigation and return a verified, cited report: plan, search broadly, check primary sources, verify adversarially, and synthesize. Run like an /autoresearch command for a deep, trustworthy answer.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: