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.
git clone --depth 1 https://github.com/brendadeeznuts1111/betting-brain-v3Wrote 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/brendadeeznuts1111/betting-brain-v3/99-floor)<a href="https://agentmods.dev/rules/brendadeeznuts1111/betting-brain-v3/99-floor"><img src="https://agentmods.dev/badge/rules/brendadeeznuts1111/betting-brain-v3/99-floor/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/rules/brendadeeznuts1111/betting-brain-v3/99-floor"><img src="https://agentmods.dev/badge/rules/brendadeeznuts1111/betting-brain-v3/99-floor.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.03067 | $0.03067 |
| Opus 5 | $0.01533 | $0.01533 |
| Sonnet 5 | $0.00613 | $0.00613 |
| Haiku 4.5 | $0.00307 | $0.00307 |
Grade A, and why
99-floor scanned grade A with 1 finding 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 12d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
✓ Health check: curl https://worker-url/health How it starts
The opening of the file, as written. The whole thing — 430 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Floor - Autonomous Operations Layer
Overview
The Floor is the self-documenting, self-healing, and self-deploying foundation layer for Forest Grove. It consolidates MCP tool usage patterns, test fixes, type handling, and deployment automation into a single source of truth.
Status: v3.3.0 Production-Ready ✅
1. MCP Tool Usage (Mandatory)
Tool Discovery
// ALWAYS use MCP tools over raw fetch
import { callMCPTool } from './src/mcp/client';
// Available tools (6 operational):
// - forest-status: Grove health monitoring
// - deploy-dashboards: Dashboard deployment
// - release: Version management (patch/minor/major)
// - live-odds: Aggregated odds (Pinnacle + Bet365)
// - live-scores: Live scores (SportsData.io)
// - push-sports-data: Analytics ingestion
Required Parameter Order
// ALWAYS follow this parameter order:
callMCPTool('place-bet', {
customerId: string, // 1. Who
eventId: string, // 2. What
stake: number, // 3. How much
odds: number, // 4. At what price
side: 'back' | 'lay', // 5. Direction
});
Idempotency
// ALWAYS pass idempotencyKey or let tool auto-generate
import { nanoid } from 'nanoid';
await callMCPTool('push-sports-data', {
data: records,
idempotencyKey: nanoid(), // Prevents duplicate processing
});
Audit Trail
// Every tool MUST write to Analytics Engine
await env.ANALYTICS_ENGINE.writeDataPoint({
blobs: [
toolName,
customerId,
eventId,
],
doubles: [stake, odds],
indexes: [`tool-${nanoid()}`],
});
2. Test-Fix Patterns (Apply Before Committing)
Error Path Tests: 429 Not 500
// ❌ WRONG: Tests expect 500 but rate limiter returns 429
expect(response.status).toBe(500);
// ✅ CORRECT: Expect 429 for rate limit responses
expect(response.status).toBe(429);
// When to use which:
// - 429: Rate limit exceeded (RATE_LIMITER KV)
// - 500: Internal server error
// - 503: Service unavailable (circuit breaker)
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.
- 12d ago First seen · 430 lines · 3,067 tokens per session scan A 553ab686b458
99-floor is a cursor rule published in the GitHub repository brendadeeznuts1111/betting-brain-v3 (8 stars, last pushed 11mo ago), licensed MIT. It adds 3,067 tokens to every session, about $0.0153 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). 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
workers-patterns
Skim this on every task that touches workers, env, routes, or bindings.
git-workflow
Cloud Agents MUST run git add, git commit, and git push before ending their response when they have made code changes. Do not end your turn without pushing. ALWAYS use Pull Requests instead of pushing directly to main.
typescript-imports
Golden rule: import each symbol from the module that defines it. A package index.ts may barrel only its own ./ modules — never another package.
new-env-var-checklist
Checklist when adding any new env secret/variable — Turbo globalEnv, GitHub sync, workflow maps, Alchemy bindings. Prevents "secret synced but worker never gets it".
dev-server
Do not start dev servers or long-running watch processes unless the user explicitly asks (e.g. “run dev”, “verify in the browser”). If they do, use the project’s bun run dev from the repo root and stop or ask before leaving a server running forever.
generated-artifacts
Agents must not manually author generated files. If generated output needs to change, edit the source-of-truth file and run the generator. If the generator fails, fix the source or generator instead of fabricating output.