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/prakrititz/relaybrain/cursorrulesgit clone --depth 1 https://github.com/prakrititz/relayBrainWhat 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 | $0.00598 | $0.00598 |
| Opus 5 | $0.00299 | $0.00299 |
| Sonnet 5 | $0.00120 | $0.00120 |
| Haiku 4.5 | $0.00060 | $0.00060 |
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 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.
This is a copy
97% identical to relay — 6 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 37 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Relay (auto-installed)
This project uses Relay for cross-agent memory. Do not ask the user to configure agent files manually.
Every session — read first: .relay/AGENT_BOOTSTRAP.md
| Phase | Action |
|---|---|
| Session start | Read .relay/relay_context.md (handoff). Never paste raw transcripts. |
| After agent work | You update .relay/*.md from compile_brief.md (stop hook triggers this) |
| Background sync | relay watch . keeps memory.json + compile_brief.md fresh |
| Not installed? | Run relay add . or relay init in project root |
Stop hooks: Cursor, Claude Code, Codex, Copilot CLI, Antigravity — installed by relay add / relay init.
Relay MCP: relay_room_brief, locks, chat history — installed into each agent's MCP config by relay add.
Relay pseudo-commands (user chat)
These are not built-in slash commands. They are a project convention — phrases the user types so agents recognize Relay intent.
When the user sends any of these phrases (with or without a leading /), treat them as Relay instructions — not casual chat:
| User says | You do |
|---|---|
/relay update or relay update |
relay sync . → relay compile . → read .relay/compile_brief.md → update .relay/project.md, current_task.md, decisions.md, failures.md → relay context . → confirm "Relay updated." |
/relay context or relay context |
Read .relay/relay_context.md (run relay context . first if stale). Summarize handoff briefly. |
/relay ask or relay ask |
Call Relay MCP tool relay_room_brief (MCP server relay). Required on Cursor, Claude Code, Codex, Copilot CLI, and Antigravity — not a terminal command for the human. Summarize teammate chat, code edits, locks, and file-sync notes. Tell the user before you edit shared files. If relay MCP is missing, say so and suggest relay add . + agent restart (do not ask the user to run relay ask). |
/relay init or relay init |
Run relay init in project root if .relay/ missing; else confirm already installed. |
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 · 37 lines · 598 tokens per session scan A 9b0701ca2e16
cursorrules is a cursor rule published in the GitHub repository prakrititz/relayBrain (31 stars, last pushed 2d ago), licensed MIT. It adds 598 tokens to every session, about $0.0030 per session on Opus 5. A static security scan graded it A with 0 findings. It is 97% identical to relay, differing in 6 lines, and is treated as a copy.
Other cursor rules, from other repositories
app_layer_rules
App layer (API) contract for FastAPI endpoints in backend/apps. Parse/validate input, call services, map domain errors to HTTP, return JSONResponse on success.
frontend_overview_rules
Frontend overview - directory structure, layer responsibilities, and dependency rules.
service_layer_rules
Service layer implements core business logic orchestration; raise custom exceptions; no HTTP handling.
hook_layer_rules
Hook layer rules for custom React hooks and state management.
page_layer_rules
Page layer rules for Next.js App Router pages and layouts.
type_layer_rules
Type layer rules for TypeScript type definitions and interfaces.