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/tw-haiven/haiven-mcp-server/cursorrulesgit clone --depth 1 https://github.com/tw-haiven/haiven-mcp-serverWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/tw-haiven/haiven-mcp-server/cursorrules)<a href="https://agentmods.dev/rules/tw-haiven/haiven-mcp-server/cursorrules"><img src="https://agentmods.dev/badge/rules/tw-haiven/haiven-mcp-server/cursorrules.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00395 | $0.00395 |
| Opus 5 | $0.00198 | $0.00198 |
| Sonnet 5 | $0.00079 | $0.00079 |
| Haiku 4.5 | $0.00040 | $0.00040 |
Grade A, and why
cursorrules 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 3d ago.
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.
What it actually says
Tests
When I am not just asking a question, but we're building a new feature, then I would like you to always use TDD:
- write a test first
- run (or let me run) the test first so it can go red (fail) and ONLY then
- write the implementation
- run the test again to see if it passes
Ask me before making decisions like choosing a technology or framework
Help me always get to the next "runnable" step that I can commit, remind me to make commits when we get to a point that is a working state
Backend
- Our backend is in Python and in the app/ subfolder. The app needs to be started from that directory.
- The backend tests are in app/tests/
- When running tests or the application, the virtual environment in app/.venv first needs to be activated
- We're using constructor dependency injections to make our code more testable. Pass dependencies in via constructors, do not initialise them inside of a class.
- When we wrap up a backend change, make sure to remind me to start up the whole application, to make sure that it's all well-integrated.
Frontend
- Our frontend is in JavaScript, React, Next, in the ui/ subfolder, code in ui/src/, tests in ui/src/tests. Make sure to navigate into the ui/ subfolder before executing yarn tasks.
- We're using Next in "client only" mode, NO server-side features, because we're distributing the frontend code as static resources that call a Python API for data
- We're using vitest for testing
- We're using the "prettier" library for formatting
- Always use remix icons for icons
- We are using yarn
- Do not create JSDocs
- Our components currently go in the ui/app folder, and the file naming convention is _name_of_component.js
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.
- 3d ago First seen · 30 lines · 395 tokens per session scan A 76b95dfd52bf
cursorrules is a cursor rule published in the GitHub repository tw-haiven/haiven-mcp-server (0 stars, last pushed 9mo ago), licensed Apache-2.0. It adds 395 tokens to every session, about $0.0020 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
tdd
Final Verification: Always use: go test -v ./backend/api/handler/... | grep FAIL.
workflow
BDD-style workflow, UI screenshots in the PR, HTTP OpenAPI and config schema sync before lint, final checks.
agent-routing.pi-input-intercept
Codex 模型經 Pi machine dispatch 提出問題時的攔截、評估、代答或升級 protocol；觸及 spectra change / screenshot 情境時 path-scoped 載入.
mcpnuke-tests
Test conventions for mcpnuke — enforces TDD workflow.
testing-discipline
TDD, BDD, and testing best practices — stack-agnostic.
onekey-test-designer
OneKey Test Designer. Invoke when user types "/onekey-test-designer", "设计用例", "写用例", "新增测试", "@ ". From PRD/recording to executable test scripts. Phase 0: Mandatory knowledge lookup via scripts/lookup-knowledge.mjs before generating any .test.mjs.