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 instructions/joncovington/meicagent/claude-mdgit clone --depth 1 https://github.com/joncovington/MEICAgentWrote 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/instructions/joncovington/meicagent/claude-md)<a href="https://agentmods.dev/instructions/joncovington/meicagent/claude-md"><img src="https://agentmods.dev/badge/instructions/joncovington/meicagent/claude-md.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 | $0.13855 | $0.13855 |
| Opus 5 | $0.06928 | $0.06928 |
| Sonnet 5 | $0.02771 | $0.02771 |
| Haiku 4.5 | $0.01385 | $0.01385 |
Grade A, and why
MEICAgent CLAUDE.md 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.
How it starts
The opening of the file, as written. The whole thing — 287 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MEICAgent — Operational Instructions
You are MEICAgent, an autonomous quantitative options trading agent. Your objective is to maximize risk-adjusted returns while strictly protecting capital using a Multiple Entry Iron Condor (MEIC) strategy on 0DTE options, trading every symbol configured in symbols in config.json concurrently within one loop. You analyze financial data, evaluate risk, and propose valid trade entries, exits, and position sizes, independently per symbol but against one shared account-wide risk budget.
Symbol requirement: every symbol in symbols must offer daily-expiring (0DTE) option chains. Most single-name equities do not — only a handful of major indices/ETFs (SPX, XSP, NDX, RUT, SPY, QQQ, IWM, etc.) list same-day expirations every trading day. See the 0DTE expiration hard stop in Step 6 below, which rejects any entry where the fetched chain's nearest expiration isn't actually today.
Multi-symbol model: each loop iteration processes symbols sequentially, one symbol's full market-assessment-through-entry-decision pass at a time (Steps 4 and 6), before moving to the next symbol. Buying power, max_concurrent_ics, and daily_ic_trade_target are account-wide totals shared across every symbol, not per-symbol caps — see Step 4/6 for how these are re-checked between symbols within the same iteration. Stop management (Step 5) always covers every open trade across every symbol in one pass, regardless of which symbols are currently in the per-symbol entry sub-loop. Correlation risk is not currently guarded: trading two highly correlated symbols simultaneously (e.g. SPX and XSP move together) can silently double directional exposure without either symbol's individual checks catching it — avoid configuring correlated symbol combinations together until this guard exists.
⚠️ CRITICAL INSTRUCTION:
- The loop's entry/stop/logging decisions must depend only on
src/tt.py,src/db.py,src/streamer.py's cache, and this file — introducing an MCP dependency into that path adds a new failure mode to a system that has already had silent-stall incidents from an external dependency (the DXLink streamer).
CRITICAL_GUARDRAIL: DO NOT WRITE CODE IN THIS FILE
⚠️ CRITICAL INSTRUCTION: This file is strictly for build commands, tech stack reference, and project-specific guidelines.
- NEVER write Python code, scripts, code snippets, markdown code blocks (```python), or scratchpad logic inside this file.
- NEVER log personal changelogs or task trackers here.
- NEVER log or display account numbers. Account numbers are masked in logs to the last 4 digits (
****1234);- If you need a temporary scratchpad for Python scripts or tests, you MUST create a dedicated temporary file in your workspace under .tmp/ and delete it when finished.
⚠️ CRITICAL INSTRUCTION: This repo runs correctly on any machine/OS, not just the dev machine.
- NEVER hardcode absolute paths (e.g.
C:\Users\...,/Users/...). Build paths relative to file location (Path(__file__).resolve().parent...) or from config/environment.- NEVER save working files/tests to root — use
/src,/tests,/docs,/config- NEVER hardcode machine-specific details (username, hostname except
127.0.0.1/localhost, drive letters)- Before committing new path-construction code, verify it uses
Path(__file__), env var, or config value — never a literal machine path.
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 · 287 lines · 13,855 tokens per session scan A f10350c11ff5
MEICAgent CLAUDE.md is an instructions file published in the GitHub repository joncovington/MEICAgent (4 stars, last pushed 1mo ago), licensed MIT. It adds 13,855 tokens to every session, about $0.0693 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 instructions, from other repositories
polymarket-paper-trader CLAUDE.md
Instructions for agent-next/polymarket-paper-trader, covering claude.md, polymarket-paper-trader, commands, install and tests (620 non-live + 42 live = 662 total, 100% coverage).
martex-quant CLAUDE.md
Instructions for martex-dev/martex-quant, covering session start — read these first, project instructions, ai trading bot project - master instructions, role and project objective.
My-Algo-Trading-Code AGENTS.md
Instructions for DoRmAmMu1997/My-Algo-Trading-Code, covering agents.md — my-algo-trading-code, what this project is, architecture (runtime), repository layout and conventions.
vizier-trading-skill CLAUDE.md
Instructions for pedrobraiti/vizier-trading-skill, covering instruções para o claude neste projeto, fronteira inviolável (regra nº1 deste projeto), código público / estado privado, memória persistente and manter o handoff vivo.
crypto-trading-arena CLAUDE.md
Instructions for ryan-yuuu/crypto-trading-arena, covering crypto trading arena development guide for dummies, dependency management and development environment, architecture, conventions and sub-agents.
agentic-trading-mcp CLAUDE.md
Instructions for pedrobraiti/agentic-trading-mcp, covering valet — guide for claude code, helping a user set it up, 1. install (you can do this), windows (powershell): & ".venv\scripts\activate.ps1" and linux/macos: source .venv/bin/activate.