martex-quant CLAUDE.md

Project instructions for an AI trading-bot codebase, including required startup documents, research rules, testing standards, and deployment criteria. The instructions require recording hypotheses, counting trials, and comparing new strategies with the deployed system.

In plain words
What is it for?
They are for guiding sessions in the trading project, recording and evaluating hypotheses, running strategy tests, maintaining research history, and deciding whether new features can be deployed.
Why use it?
They give the agent the project’s current state and research history so it does not repeat rejected ideas or make unsupported trading claims. They also set consistent rules for deciding whether a strategy is ready for deployment.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/martex-dev/martex-quant/claude-md
Clone the repo
git clone --depth 1 https://github.com/martex-dev/martex-quant
Per session 6,161 This file is loaded in full into every session.
When invoked 6,161 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.06161 $0.06161
Opus 5 $0.03080 $0.03080
Sonnet 5 $0.01232 $0.01232
Haiku 4.5 $0.00616 $0.00616

Measured 2d ago against content hash 790ad40fda76, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

martex-quant 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 2d 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.

CLAUDE.md · 633 lines

How it starts

The opening of the file, as written. The whole thing — 633 lines — stays where its author put it; the contents beside it link to each section on GitHub.

SESSION START — READ THESE FIRST

  1. PROJECT_STATE.md — what is running right now, schedules, the firm, the eval plan, next actions.
  2. PROJECT_MEMORY.md — full hypothesis ledger (105 trials), all results, meta-findings, lessons, open caveats.

Do not re-derive or re-litigate anything recorded there. Do not re-test killed ideas without a new pre-registered spec and a stated reason.

NON-NEGOTIABLE OPERATING RULES (distilled from 23 hypotheses)

  • Pre-register every hypothesis: numbered doc in docs/hypotheses/ with verdict bars, COMMITTED BEFORE any test runs. Count every trial (including variants and descriptive horizons) in the ledger; DSR is always benchmarked against ALL trials ever run.
  • Kill test (cheap information study) BEFORE any strategy build. Event-driven engine is the source of truth for strategies; vectorized screening only pre-engine.
  • New features must beat the DEPLOYED system incrementally, not zero — for DEPLOYMENT decisions. Amended 2026-08-27 (docs/research/standalone-viable-amendment.md): a hypothesis that clears the full standalone bar (positive after costs, CI excluding zero, DSR_global >= 0.95, engine-grade) but does NOT beat the incumbent is closed STANDALONE-VIABLE, not KILLED. It is not deployed; it is a live edge on the bench. Existence and deployment are different questions and the incremental bar only answers the second. No retroactive relabelling: old KILLED verdicts stand until re-registered and re-run.
  • Paper accounts run only validated/eligible specs; one spec per record (spec change = archive the record, fresh $5,000 start).
  • Live/real-money actions are gated: the runbook (docs/research/eval-runbook.md) governs the eval; going live is a deliberate CLI step, never a dashboard button; the guard's KILLED latch is cleared only by a human.
  • Every session: run pytest + ruff + strict mypy before committing; commit per component; push to origin (GitHub martex-dev/martex-quant).
  • Restart the dashboard server after changing dashboard code.
  • Report negative results with the same rigor as positive ones. The ledger's honesty is the project's only real asset.

Read the full file on GitHub · 633 lines

Changes

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.

  1. 2d ago First seen · 633 lines · 6,161 tokens per session scan A 790ad40fda76

Subscribe to this mod's changes

martex-quant CLAUDE.md is an instructions file published in the GitHub repository martex-dev/martex-quant (20 stars, last pushed 5d ago), licensed MIT. It adds 6,161 tokens to every session, about $0.0308 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 2026-08-30.

Related

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).

agent-next/polymarket-paper-trader · 1,257 tokens

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.

DoRmAmMu1997/My-Algo-Trading-Code · 4,730 tokens

lumibot CLAUDE.md

Claude Code instructions for Lumiwealth/lumibot, covering claude.md - ai assistant instructions for lumibot, public repo secret hygiene (critical), 🚨🚨🚨 rule #1 — never fabricate backtest data 🚨🚨🚨, quick start and image generation rule (critical).

Lumiwealth/lumibot · 8,606 tokens

trading-ops CLAUDE.md

Instructions for l3lackcurtains/trading-ops, covering trading-ops — project rules, authoritative protocols, hard boundaries (never cross), self-improvement loop (applies to every command) and automation scripts.

l3lackcurtains/trading-ops · 2,279 tokens

My-Algo-Trading-Code CLAUDE.md

Instructions for DoRmAmMu1997/My-Algo-Trading-Code, covering claude.md — my-algo-trading-code, what this project is, architecture (runtime), repository layout and conventions.

DoRmAmMu1997/My-Algo-Trading-Code · 4,740 tokens

QuantDesk CLAUDE.md

Claude Code instructions for 0xbet-ai/QuantDesk, covering claude.md, purpose, read this first, dev setup and environment variables.

0xbet-ai/QuantDesk · 2,018 tokens