10x
The engineering discipline that makes your AI developer 10x — hill-climbed through billions of tokens of autonomous experimentation.
Install
npx skills add z3z1ma/10xclaude mcp add 10x -- npx skills add z3z1ma/10x{
"mcpServers": {
"10x": {
"command": "npx",
"args": [
"skills",
"add",
"z3z1ma/10x"
]
}
}
}Install command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →10x
An opinionated instruction set that makes AI coding agents clarify, remember, verify, and improve across sessions.
10x is a drop-in Markdown instruction set for AI coding agents.
It is one self-contained file: SKILL.md. Load it into Codex,
OpenCode, Claude Code, Cursor, Gemini CLI, or any harness that reads project
instructions. It teaches the agent to:
- challenge vague work before coding;
- preserve repo-local context in
.10x/; - split discovery from execution;
- prove changes with evidence and review;
- turn mistakes, dead ends, and repeated friction into reusable knowledge.
Try it in 60 seconds, either way:
Skills CLI
npx skills add z3z1ma/10x
Copy-paste
Open SKILL.md and append everything after the YAML frontmatter to
whatever instruction file your agent already reads.
Both paths are first-class. The Vercel Skills CLI
handles placement across 70+ agent harnesses; copy-paste is maximally flexible
because the product surface is still just Markdown: one SKILL.md, plain
English, easy to inspect, fork, edit, splice, or paste. No runtime, no service,
no database.
Read the full instructions | Install | How it is tested
Why this exists
You've worked with a 10x developer. They are not 10x because they type faster. They are 10x because they eliminate entire classes of problems, write less code, record why decisions were made, and leave the next engineer further ahead than they started.
Most AI coding agents live at the opposite end of that spectrum. They are strong at local implementation, but weak at durable judgment: carrying context across sessions, challenging vague requirements, proving that "done" means done, and remembering what they learned after the chat closes.
10x gives the agent a working method. It searches existing context before asking you to repeat yourself, pushes back when a request would require invented semantics, breaks work into bounded tickets, captures observed evidence instead of vibes, and runs a retrospective so the next session inherits the lesson.
Thursday's agent uses Tuesday's judgment because Tuesday's agent worked carefully enough to leave a trail.
This is my personal instruction set - the base instructions I load into every coding harness I use. It is opinionated, plain-text, and portable.
What still gets lost
Maybe your workflow already has good parts: plan mode, spec files, custom skills, subagents reviewing subagents. The pieces still often operate like a ticket machine - accepting work at face value, executing in isolation, and declaring victory without verification.
| You already use | What still goes wrong |
|---|---|
| Plan mode | Reasoning disappears with the session |
PLAN.md or spec files |
Plans rarely link to evidence, reviews, or later decisions |
| Subagents | Intermediate reasoning, checkpoints, and validation vanish into a final reply |
| Custom skills | They use different state and vocabulary from each other |
| "Think before coding" prompts | The agent still guesses when requirements are vague |
| Chat history | Important conclusions stay implicit, private, or impossible to grep |
The gap is not tooling. It is judgment. 10x gives those failures an explicit place to be caught and turns the prevention mechanism into durable project context.
How it works
Similar plugs
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local…
Why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.
Production-grade engineering skills for AI coding agents.
The 1st Collaborative AI Design Agent Workspace. Local-first and open source. Create prototypes, websites, apps, dashboards, presentations, images, and videos with your team…