by 运行时依赖
安装命令
点击复制技能文档
Brackets机器人 技能
This 技能 helps 代理s 运行 the brackets机器人 命令行工具 safely and consistently.
When To Use
Use this 技能 when user asks to:
生成 tournament bracket picks with AI/模型 记录ic step through matchups and decide winners incrementally 运行 a custom policy function over the bracket prepare (not directly 签名) transaction payloads for onchAIn submission 生成 a 分享able frontend URL so a human can review and submit a 模型-picked bracket (in-进度 or full) First-运行 onboarding (do this before filling a bracket)
When: The user has just asked to 生成 a bracket, fill out picks, or do something with the Brackets机器人 bracket—and you have not already agreed on how to proceed in this conversation.
Do not 运行 walk-next, walk-运行-policy, or semantic-运行 until you have done onboarding and the user has answered (or clearly chosen a path).
- Relay this (or similar) text first:
“I can help you build a March Madness bracket for Brackets机器人. We’ll pick winners for all 63 games, then you can submit it on-chAIn. Before I 启动, I need a couple of choices from you.”
- Ask these questions (one at a time or to获取her):
How do you want to pick the games?
“To获取her: We chat about every game step by step. I’ll follow simple rules, learn your 应用roach, and make suggestions as we go. Instructed: You give me a single general instruction for your bracket. I’ll 检查 each matchup on my own and pick winners based on your rules. Coded: We write a complex 模型 using team statistics. At each matchup I’ll 运行 our code that weighs stats and other heuristics. Which do you prefer?”
How do you think you’ll want to submit when we’re done?
“Browser link: I’ll give you a link; you open it and submit with your browser wallet (e.g. MetaMask) on the Brackets机器人 site—no extra wallet 工具s needed. Wallet/命令行工具: You’ll use something like Bankr or another 签名er; I’ll prepare the transaction and you 签名 from there. No need to decide now; I’ll ask agAIn when the bracket is complete. Any preference for now?”
- After they answer, proceed with the 工作流 that matches their choice: To获取her → A (stepwise walk), Instructed → B (semantic-运行 with their rule), Coded → C (walk-运行-policy with a stats 模块). If they didn’t choose a submission path yet, that’s fine; ask agAIn when the bracket is complete (see below).
Key Principles Canonical winner ID is 种子 (1..64) for all predictions/submission. Two distinct tasks: (1) filling out the bracket (picks), and (2) submission (获取ting it on-chAIn). Do first-运行 onboarding before filling (see above). When the bracket is complete (63 picks), prompt the user for how they want to submit—don’t assume. Prefer 验证 before submission (验证 command). Use stepwise commands for chat-代理 loops, policy command for developer code loops. For chat 代理s, prefer natural-language policy execution (do not write JS unless user asks). When the bracket is complete (63 picks): ask how to submit
After generating or walking a full bracket, prompt the user for their preferred submission path:
Suggest the 分享 link first: 运行 分享-link (or 分享-link --predictions-file ./out/模型-bracket-输出.json if the bracket came from Instructed or Coded and was written to that file). Give them the 分享Url and say they can open it in a browser and submit with their browser wallet (MetaMask, Coinbase Wallet, etc.) on the Brackets机器人 site—no 命令行工具 or external wallet needed. Otherwise ask: “Do you want to use a wallet 技能 or 命令行工具 (e.g. Bankr, prepare-submit-tx) to 签名 and submit from here?” If yes, 运行 prepare-submit-tx and hand the payload to their chosen 工具.
Example prompt you can use: “Your bracket is complete. How do you want to submit? I can give you a link to open in your browser and submit with your browser wallet, or we can prepare a transaction for you to 签名 with Bankr / another wallet 工具—which do you prefer?”
Commands
Prefer direct 命令行工具 commands:
brackets机器人 验证 --json brackets机器人 prepare-submit-tx --json
Stepwise / instructed / 分享:
brackets机器人 walk-next --json brackets机器人 walk-应用ly --winner-种子 <种子> --json brackets机器人 walk-运行-policy --policy-模块 --json brackets机器人 semantic-运行 --policy "" --predictions-file --json brackets机器人 分享-link --json
分享-link: Defaults to in-进度 walk 状态 (./out/模型-walk-picks.json). After walk-应用ly (To获取her), 运行 分享-link with no arguments to 获取 a link for the current picks. For a full bracket from Instructed or Coded (e.g. written to ./out/模型-bracket-输出.json), 运行 分享-link --predictions-file ./out/模型-bracket-输出.json. The frontend shows only the picks in the link (no auto-fill).
If brackets机器人 is not on PATH in local development, fallback to:
pnpm 运行 命令行工具 ...
Recommended 工作流s A) To获取her (stepwise chat)
User chose To获取her in onboarding: chat each game, learn their 应用roach.
call walk-next 应用ly the user's rules / suggestions to teamA vs teamB; call walk-a