Prop Firm Sim MCP
Monte Carlo simulation of prop-firm challenges over each firm's exact rules: pass odds, expected cost, EV and risk per trade.
- 110 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Results influence financial decisions: they are probabilities under assumptions, not a guarantee
- For firm data the server calls the public LuxAlgo directory
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/prop-firm-simDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Results influence financial decisions: they are probabilities under assumptions, not a guarantee
- For firm data the server calls the public LuxAlgo directory
This is third-party code. Review the repository files before installing.
What it does
Prop Firm Sim plays ten thousand complete paths through a prop-firm challenge ruleset based on your trading statistics or actual trade history. The output is the pass probability for each phase and overall with confidence intervals, expected attempts, total cost, expected value with every fee priced in, time to funding and the chance that a funded account collects a payout. The engine is pure deterministic TypeScript: the same seed gives byte-identical numbers. The project consists of a core, a CLI and an MCP server. Through MCP an agent gets the tools `list_firms`, `get_challenge_rules`, `simulate_challenge`, `optimal_risk`, `compare_challenges`, `bootstrap_simulate` and `analyze_portfolio_overlap`. Firm rules come from LuxAlgo's public directory without a key, and any ruleset can be passed inline for fully offline work. The tool places no trades.
Who it is for. Traders and analysts who want to see the odds in numbers before paying for a prop-firm challenge.
Good fit when
- Before buying a challenge the pass probability and expected cost of attempts are estimated.
- Several challenges are compared by EV for your own statistics.
- Risk per trade is chosen for the highest pass chance or the highest EV.
- Actual trade history in R multiples is run through a firm ruleset.
Not a fit when
- An investment recommendation or a guaranteed result is wanted: the tool is not an advisor.
- You want to trade automatically: it sends no orders.
- Firm rules are vague: the tool refuses to guess ambiguous clauses.
Example request
Simulate the FTMO 100K 2-step challenge at a 48% win rate, 1.6R average win and 1R loss, 1% risk, and show the pass probability and EV.MCP
- Transport
- stdio, http
- Authentication
- not required
Environment variables
LUXALGO_APP_ORIGIN- Optional firm directory origin instead of the public LuxAlgo address.
Limitations
Every result is a distribution under stated assumptions rather than a promise, and the tool response carries assumption flags, data provenance and a disclaimer that the agent must show. The model relies on your statistics and may diverge from real trading. For firm data the server calls LuxAlgo's public directory and all other calculations run locally without telemetry. Check rule and price questions on the firm's own site. The repository was last updated on 28 October 2026 and uses the MIT licence.
How to disable. Run `claude mcp remove prop-firm-sim` or delete the entry from the client configuration. The server stores no local data.
FAQ
Does the tool place trades?
No. It only simulates probabilities from your statistics and challenge rules, with no orders or payments.
Where do firm rules come from?
From LuxAlgo's public directory without a key. Any rules can be passed inline through `spec` for work without a network.
Can actual trade history be used?
Yes, `bootstrap_simulate` takes trades in R multiples and accounts for real losing streaks with a block bootstrap.
Are numbers reproducible?
Yes, the engine is deterministic: the same seed with the same inputs gives an identical result.
Is there telemetry?
No, none in any package. The only outbound call is fetching the firm directory.
Discussion
No questions or reviews yet.
Related
MCP serversMarket data in a dialogue with the agent: hosted server or local run, three modes and tool sets loaded on demand
A committee of four AI roles reviews a portfolio, keeps a decision log and runs backtests with lookahead protection
An MCP server for agentic trading: backtesting, 462 prebuilt alphas, trade journal analysis and multi-agent research teams
A financial analysis MCP server: 500+ methods from company statements, ratios, indicators, risk metrics and macro data
A personal family office in the terminal: typed portfolio calculators, a local ledger and agents for Claude Code and Codex.
One MCP server combines an indicator encyclopedia, a trade journal, broker account reading and US market data, no orders.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Prop Firm Sim (this card) | yes | not required | 110 | |
| PlainBacktest | no | 529 | backtests trading strategies on history; here the odds of passing a prop-firm challenge under its rules are simulated | |
| Claude Trading Skills | no | 3,800 | skills for trading tasks with an agent; here a dedicated calculation of probability, attempt cost and EV under firm rules |
README in short
The README takes an honest tone: it shows a console calculation example, explains why odds are worse than expected and links the rule semantics and disclaimer.