Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add rules/maccman/agent-playground/project-guidelinesgit clone --depth 1 https://github.com/maccman/agent-playgroundWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00677 | $0.00677 |
| Opus 5 | $0.00338 | $0.00338 |
| Sonnet 5 | $0.00135 | $0.00135 |
| Haiku 4.5 | $0.00068 | $0.00068 |
Grade A, and why
project-guidelines scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Playground - Project Guidelines
Project Overview
This is an AI agent playground for generating autonomous scripts and workflows. All agents are designed to perform complex, multi-step tasks including web automation, bookings, data research, and administrative functions.
Agent Development Process
- Planning First: Always come up with a good plan before executing any task
- Write and run typescript or python code: Write and run typescript/python code to test your ideas. For anything that is best done programmatically, like math or other types of logic, then just write a quick TypeScript/python program and run it.
- ALWAYS prefer running executables vs writing scripts: Running executables is faster than writing scripts.
- Data from the web: Use the
web_searchandmcp_browsermcptools to get data from the web. Never use themcp_browsermcptool to Google search - you should use the dedicatedweb_searchtool for that as it's more reliable. - Use unix principles: Use unix principles to build your scripts. Use
stdinandstdout. Use|to chain commands together. - Stream output: Stream output to the console. e.g. when processing CSVs, output line by line.
- Never alter a file in the
data/directory in-place: Always either produce a copy or, if you're editing textual data, usestdinandstdoutto edit the data.
AI SDK
To communicate with AI providers, use the Vercel AI SDK.
- Use Vercel AI SDK: Use the AI SDK directly with providers like
@ai-sdk/openai,@ai-sdk/anthropic,@ai-sdk/google - No Custom Abstractions: Don't create custom AI client wrappers - use the SDK directly
- Environment variables for AI providers are automatically handled by Vercel AI SDK
- No need to manually check or validate API keys in code
- Standard variables:
OPENAI_API_KEY,ANTHROPIC_API_KEY,GOOGLE_GENERATIVE_AI_API_KEY
Project Structure
lib/- Reusable utilities and shared codescripts/- Runnable task scripts (use tsx for execution)types/- Shared TypeScript type definitionsdata/- Storage for data files (Markdown and CSV)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 59 lines · 677 tokens per session scan A 6f80c5740c3d
project-guidelines is a cursor rule published in the GitHub repository maccman/agent-playground (2 stars, last pushed 6mo ago), licensed MIT. It adds 677 tokens to every session, about $0.0034 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other cursor rules, from other repositories
webkit-browser
Cursor rule "webkit-browser" from duckduckgo/apple-browsers, covering webkit & browser development guidelines, webview configuration, basic webview setup, user scripts management and tab management.
cypress-e2e-testing-cursorrules-prompt-file
Cursor rules for Cypress development with E2E testing.
vasu-playwright-utils
../../templates/cursor-rules/vasu-playwright-utils.mdc.
dev-browser
Fallback browser automation with persistent Chrome state. Use only when Browser Use is unavailable or blocked.
node-dependencies
Enforce Node.js versioning and package management best practices.
security-standards
Cursor rule "security-standards" from wjgogogo/cursor-rules, covering 安全规范, 核心原则 [p0], 输入验证, xss 防护 and csrf 防护.