{
  "markdown": "<img src=\"https://raw.githubusercontent.com/kellyruns/kellyruns/main/assets/hero.svg\" alt=\"Kelly — autonomous trading agent on Robinhood Chain. Nobody pays to be wrong.\" width=\"880\" />\n\n<p align=\"center\">\n  <a href=\"https://kellyruns.xyz\"><b>kellyruns.xyz</b></a> ·\n  <a href=\"https://kellyruns.xyz/how-it-works\">how it works</a> ·\n  <a href=\"https://x.com/KellyRunsIt\">@KellyRunsIt</a>\n</p>\n\n<br />\n\n## ◢ whoami\n\n```ts\nconst kelly = {\n  what:      \"autonomous trading agent that shows its work\",\n  chain:     \"Robinhood Chain · 4663 · Arbitrum stack\",\n  loop:      [\"observe\", \"think\", \"decide\", \"act\", \"learn\", \"broadcast\"],\n  publishes: \"the reasoning behind every call — including the losing ones\",\n  ledger:    \"append-only. failures and rules-engine rejections included\",\n  execution: \"paper mode. the EVM path is wired and deliberately not armed\",\n  thesis:    \"nobody pays to be wrong\",\n} as const;\n```\n\n<br />\n\n<img src=\"https://raw.githubusercontent.com/kellyruns/kellyruns/main/assets/card.svg\" alt=\"Kelly status card: paper mode, no vault deployed, sizing against the 5th percentile.\" width=\"880\" />\n\n<br />\n\n## ◢ the strategy lab\n\nMost people learn leverage, liquidation and impermanent loss by losing money to\nthem. The lab is the cheaper way round.\n\nYou submit a lowcap, perp or LP strategy as **parameters, not prose**. The lab\nstress-tests it across thousands of simulated paths and publishes the **whole\ndistribution** — not one backtest number that hides the tail where the damage\nlives.\n\n| | |\n| --- | --- |\n| Certification gate | **≥ 3 runs** and **≥ 1,500 paths** |\n| How to size | against the **5th percentile** the run prints, never the median |\n| What a pass means | a statement about a distribution — **not a prediction** |\n| Who pays | trading fees cover the $0.85 of compute per run. Not you. |\n\n<br />\n\n## ◢ connect an agent\n\nKelly speaks MCP, so another agent can read the record without scraping HTML —\nand the reproduction seed comes with it.\n\n```bash\nclaude mcp add --transport http kelly https://kellyruns.xyz/mcp\n```\n\nStateless streamable HTTP, protocol `2025-06-18`, no auth and no key. Published\nto the official registry as\n[`xyz.kellyruns/strategy-lab`](https://registry.modelcontextprotocol.io/v0/servers?search=kellyruns).\n\n**Read-only by design.** No tool here submits a strategy, funds a run, or moves\nanything — writes belong behind a wallet session, and an unauthenticated\nendpoint is the wrong place to re-implement one.\n\n## Tools\n\n| Tool | What it returns |\n| --- | --- |\n| `get_agent_state` | Treasury, open positions, recent trades, and the reasoning behind the latest cycle. Paper mode — a published track record, not a portfolio. |\n| `get_certification_gate` | The thresholds a strategy must clear, and what a pass does and does not mean. Read this first. |\n| `get_lab_summary` | Counts by status, testing-pool balance, and how many runs it can still fund. |\n| `list_strategies` | Strategies filtered by kind and status, each with its certification rollup — runs, paths, median PnL, worst case, ruin rate. |\n| `get_strategy` | One strategy in full: parameters, verdict, every check the gate ran, and each run's PRNG seed. |\n\nThe seed is the point. Re-running with it reproduces the result exactly, which\nis what makes a verdict something you can disagree with by running it rather\nthan by arguing about it.\n\n<br />\n\n## ◢ why \"Kelly\"\n\nThe **Kelly Criterion** (John L. Kelly Jr., Bell Labs, 1956) answers the question\nretail traders skip: not *is this a good bet* but ***how much should I bet***.\nThat is the direct ancestor of the risk clamp that gets the last word on sizing\nhere, over anything the model proposed.\n\nWhich is also why the agent is a character and not a dashboard. `Kelly says no`\nis a sentence. A tool can't be the subject of one.\n\n<br />\n\n## ◢ honest status\n\nUp front, because a project about publishing losses cannot be coy about its own\nreadiness.\n\n```\nexecution   paper mode. every fill simulated, slippage and fees modelled\n            pessimistically so edge isn't an artifact of frictionless fills\nevm path    wired, deliberately NOT armed — pending real quoting, bounded\n            approvals, a token safety check, and nonce management\nvault       no contract deployed. /vault shows a requirements checklist,\n            not a deposit button. nothing here custodies anyone's funds\nwallet      connecting signs an identity statement and nothing else.\n            no transaction, no approval, no transfer\nadvice      none of this is financial advice. the agent is not a licensed\n            advisor, and a certification is not a prediction\n```\n\n<br />\n\n## ◢ on chain\n\n| | |\n| --- | --- |\n| Chain | Robinhood Chain — `4663`, Arbitrum stack, ETH gas |\n| ERC-8004 Identity Registry | [`0x8004A169…a432`](https://robinhoodchain.blockscout.com/address/0x8004A169FB4a3325136EB29fA0ceB6D2e539a432) |\n| Agent registration | [`/.well-known/agent-registration.json`](https://kellyruns.xyz/.well-known/agent-registration.json) |\n\n<br />\n\n## ◢ built with\n\n`TypeScript` · `Next.js` · `viem` · `Postgres` · `Anthropic SDK`\n\nThe art above is generated by [`scripts/build-art.mjs`](https://github.com/kellyruns/kellyruns/blob/main/scripts/build-art.mjs) —\ntwo self-contained SVGs, no external fonts and no remote assets, because\nGitHub's image proxy blocks every outbound fetch. No third-party widget\nservices either, so nothing here breaks when someone else's server goes down.\n\n---\n\n<sub>Built on Robinhood Chain. Not affiliated with, endorsed by, sponsored by, or\npartnered with Robinhood Markets, Inc. The chain is permissionless — anyone can\ndeploy on it without approval, and that includes this project.</sub>\n",
  "bytes": 5673,
  "sha": "ebc12102d72ac508eb87be63fafe17c84d05af185b3e81c4bf3ec92d6d27f1cb",
  "repo_slug": "kellyruns/kellyruns",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_xyz_kellyruns_strategy_lab_2df86a99/readme"
}