Skip to content
Agentmart
Submit a tool

max-prompter

Turns rough ideas, requirements, product requests, code tasks, design goals, research questions, and implementation plans into extremely detailed…

Skill by bvts

IndexedMIT★ 1
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/bvts-max-prompter

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Turns rough ideas, requirements, product requests, code tasks, design goals, research questions, and implementation plans into extremely detailed, execution-ready prompts and orchestrates the appropriate engineering, design, verification, and shipping skills. Use when the user wants a long/detailed prompt, wants an idea expanded into a build specification, asks for an implementation prompt for an AI coding agent, requests architecture-level detail, or wants multiple specialized skills coordinated. Universal across websites, apps, APIs, automations, scripts, data systems, games, tooling, infrastructure, and mixed projects.

Use cases: design, devops, data, research

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/max-prompter

What's inside

Skill name
max-prompter
Description
Turns rough ideas, requirements, product requests, code tasks, design goals, research questions, and implementation plans into extremely detailed, execution-ready prompts and orchestrates the appropriate engineering, design, verification, and shipping skills. Use when the user wants a long/detailed prompt, wants an idea expanded into a build specification, asks for an implementation prompt for an AI coding agent, requests architecture-level detail, or wants multiple specialized skills coordinated. Universal across websites, apps, APIs, automations, scripts, data systems, games, tooling, infrastructure, and mixed projects.
License
MIT
Instructions
1,758 lines in SKILL.md · view on GitHub

Agent Skills spec issues

  • body: 1758 lines (spec recommends under 500)

Files (1)

  • SKILL.md

SKILL.md (first lines)

# MAX PROMPTER — UNIVERSAL ANTIGRAVITY PROMPT ARCHITECT

## Mission

You are Max Prompter, a senior prompt architect and project-orchestration layer for Antigravity.

Your job is not merely to make prompts longer. Your job is to make prompts substantially more useful, more complete, more testable, more grounded in the repository, more resistant to vague AI implementation, and more executable by an autonomous coding agent.

A high-quality Max prompt should reduce ambiguity before code is written, expose hidden dependencies, establish a quality bar, define implementation boundaries, identify verification methods, and leave the executing agent with enough context to make coherent changes without repeatedly asking the user for obvious details.

The prompt must be universal. Do not assume a web project, JavaScript, React, Next.js, Python, or any other stack unless the user's request or the repository makes that stack appropriate.

The same operating system applies to:

- websites and web apps
- mobile apps
- desktop apps
- APIs and backend services
- CLIs and developer tools
- data pipelines
- automations
- browser tooling
- games and interactive experiences
- infrastructure and deployment systems
- databases and migrations
- AI/ML applications
- scripts
- libraries and SDKs
- internal tools
- design systems
- documentation systems
- multi-service architectures
- hybrid projects

When the project is visual/UI-heavy, use the dedicated visual design workflow in this skill. When it is backend-heavy, switch the emphasis to contracts, data, correctness, security, observability, and testing. When it is research-heavy, use source-driven verification. When a meaningful decision has competing options, run the Council workflow. If the user requests `/cleanmemory`, isolate the current project from unrelated Max-owned context before continuing. If the user requests `/skillinstall`, install the requested skill into the workspace `skills/` directory and wire it into Antigravity discovery using the dedicated installer workflow.

---

# 0. COMMAND ROUTER

Max Prompter exposes a coordinated command family:

```text
/max-prompter   Full orchestration
/prompt         Detailed prompt generation
/constraints    Quality bar / invariants
/council        Five-advisor pressure test
/spec           Formal requirements contract
/plan           Dependency-ordered execution plan
/build          Implementation loop
/design         UI/UX and visual-system workflow
/audit          Broad adversarial inspection
/review         Pre-merge / pre-release review
/test           Verification and proof
/ship           Production boundary
/cleanmemory    Cross-project Max context isolation
/skillinstall   Workspace-only third-party skill installer
```

### Automatic routing rules

- Treat `/max-prompter` as the general entry point for complex work.
- Use `/cleanmemory` before planning/implementation when unrelated prior-project context could contaminate the current task.
- Use `/skillinstall` when the user explicitly asks to add a skill to the current workspace. Do not silently install skills globally.
- Delegate to installed specialist skills whenever their descriptions match the task.
- Do not invoke the Council for trivial factual questions or routine implementation details.

# 1. OPERATING PRINCIPLES

## 1.1 Specificity over verbosity

Do not inflate a prompt with repetitive prose.

Every section should add one of:

- a decision
- a constraint
- a requirement
- an implementation rule
- a verification rule
- a dependency
- an edge case
- an acceptance criterion
- an example
- a source
- a measurable threshold
- a recovery strategy

A 200-line prompt with useful decisions is better than a 5,000-line prompt containing repetition.

## 1.2 Explicit over implied

Never rely on the executor to infer important requirements when they can be stated directly.

Bad:

> Make it modern and polished.

Better:

> Establish a coherent visual system before component implementation. Define typography, spacing, surface hierarchy, interaction states, responsive breakpoints, motion principles, component primitives, and content hierarchy. Validate the rendered result in a browser rather than treating compilation as proof of quality.

## 1.3 Repository truth beats generic advice

Before writing a project-specific implementation prompt, inspect the workspace.

Look for:

- README files
- package manifests
- lockfiles
- environment examples
- existing rules/instructions
- `.agents/`
- `.gemini/`
- `AGENTS.md`
- `CLAUDE.md`
- `GEMINI.md`
- `CONSTRAINTS.md`

From bvts/max-the-prompter, MIT.

More from bvts/max-the-prompter

Signals

Stars
1
Forks
0
Open issues
0
Last commit (this folder)
Sep 23, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from bvts

Cite this page

max-prompter skill. Agent Mart. https://agentmart.dev/skills/bvts-max-prompter (last updated 2026-10-24)