← 返回
AI智能 Key 中文

Polymarket Ai Tech Trader

Trades Polymarket prediction markets on AI model releases, tech IPOs, product launches, GPU infrastructure milestones, and AI regulation events. Use when you...
在Polymarket预测市场上交易AI模型发布、科技IPO、产品发布、GPU基础设施里程碑以及AI监管事件的预测。
diagnostikon
AI智能 clawhub v0.0.3 3 版本 99857.1 Key: 需要
★ 0
Stars
📥 699
下载
💾 5
安装
3
版本
#latest

概述

AI & Tech Launch Trader

> This is a template.

> The default signal is keyword discovery + LMSYS Chatbot Arena leaderboard monitoring — remix it with AI benchmark APIs (MMLU, HumanEval), tech news RSS feeds, SEC EDGAR filings for IPO signals, or GitHub commit activity as model release early warning.

> The skill handles all the plumbing (market discovery, trade execution, safeguards). Your agent provides the alpha.

Strategy Overview

AI markets are the fastest-growing category on Polymarket in 2026. AI investment drove >90% of US GDP gains in H1 2025. This skill trades:

  • AI model benchmarks — Which company leads LMSYS rankings at a given date
  • Model releases — GPT-5, Claude 4, Gemini Ultra release/performance questions
  • Tech IPOs — OpenAI, Databricks, Stripe IPO announcement markets
  • Product launches — Apple Vision Pro 2, Tesla FSD milestones, Tesla Optimus
  • AI regulation — EU AI Act enforcement, US federal AI legislation
  • Infrastructure — NVIDIA datacenter revenue, H100 deployment counts

Key insight: AI news cycles are fast and retail trades on headlines. Informed traders with benchmark data have significant edge.

Signal Logic

Default Signal: Benchmark Divergence + News Catalyst

  1. Discover active AI/tech markets on Polymarket
  2. Monitor LMSYS Chatbot Arena for ranking shifts
  3. Monitor Hugging Face Open LLM Leaderboard for benchmark updates
  4. Compare quantitative model performance data vs market implied probability
  5. When a model clearly leads benchmarks but market hasn't repriced, enter
  6. For IPO markets: monitor SEC Form S-1 EDGAR filings as leading indicator

Remix Ideas

  • GitHub API: Watch for sudden commit activity/new repo creation on known model orgs
  • Perplexity/Google Trends: Rising AI search terms as momentum signal
  • HuggingFace API: Model download counts as proxy for adoption
  • SEC EDGAR: Automated S-1 filing alerts for IPO markets
  • NVIDIA earnings call transcripts: Forward guidance vs market pricing on infrastructure markets

Market Categories Tracked

AI_TECH_KEYWORDS = [
    "AI", "GPT", "Claude", "Gemini", "OpenAI", "Anthropic", "Google",
    "model", "benchmark", "AGI", "ChatGPT", "LLM",
    "IPO", "valuation", "Stripe", "Databricks", "SpaceX IPO",
    "Apple", "Vision Pro", "Tesla FSD", "Optimus", "robot",
    "NVIDIA", "GPU", "H100", "datacenter", "quantum",
    "EU AI Act", "regulation", "congress"
]

Risk Parameters

ParameterDefaultNotes
---------------------------
Max position size$40 USDCHigh-liquidity markets allow larger size
Min market volume$10,000AI markets are deeply liquid
Max bid-ask spread6%Tight spreads expected in popular markets
Min days to resolution5News cycles move fast
Max open positions10AI/tech is broad category

Edge Opportunities

Retail Recency Bias

After a major AI announcement (e.g., GPT-5 release), retail traders overweight recency and push other company markets down more than warranted. Fade the overreaction on competing platforms.

Benchmark Lag

Markets often price on last-known model rankings. When a new benchmark publishes at ~midnight UTC, markets take 1–4 hours to fully reprice. Tight monitoring loop captures this window.

Installation & Setup

clawhub install polymarket-ai-tech-trader

Requires: SIMMER_API_KEY environment variable.

Cron Schedule

Runs every 5 minutes (/5 *). AI news breaks around the clock; tightest loop of all category skills.

Safety & Execution Mode

The skill defaults to paper trading (venue="sim"). Real trades only execute when --live is passed explicitly.

ScenarioModeFinancial risk
--------------------------------
python trader.pyPaper (sim)None
Cron / automatonPaper (sim)None
python trader.py --liveLive (polymarket)Real USDC

The automaton cron is set to null — it does not run on a schedule until you configure it in the Simmer UI. autostart: false means it won't start automatically on install.

Required Credentials

VariableRequiredNotes
---------------------------
SIMMER_API_KEYYesTrading authority — keep this credential private. Do not place a live-capable key in any environment where automated code could call --live.

Tunables (Risk Parameters)

All risk parameters are declared in clawhub.json as tunables and adjustable from the Simmer UI without code changes. They use SIMMER_-prefixed env vars so apply_skill_config() can load them securely.

VariableDefaultPurpose
----------------------------
SIMMER_MAX_POSITION40Max USDC per trade (full conviction)
SIMMER_MIN_TRADE5Min USDC per trade (weak conviction floor)
SIMMER_MIN_VOLUME10000Min market volume filter (USD)
SIMMER_MAX_SPREAD0.06Max bid-ask spread (0.10 = 10%)
SIMMER_MIN_DAYS5Min days until market resolves
SIMMER_MAX_POSITIONS10Max concurrent open positions
SIMMER_YES_THRESHOLD0.38Buy YES if market probability ≤ this value
SIMMER_NO_THRESHOLD0.62Sell NO if market probability ≥ this value

Conviction-based position sizing

Position size scales automatically with signal strength — no flat bets.

  • At the threshold boundary (e.g. p=38% for YES): minimum trade (SIMMER_MIN_TRADE, default $5)
  • At maximum conviction (p=0% for YES, p=100% for NO): full position (SIMMER_MAX_POSITION, default $40)
  • Everything in between scales linearly

Example YES trades:

Market probabilityConvictionSize
--------------------------------------
38% (at threshold)0%$5 (floor)
30%21%$8
20%47%$19
5%87%$35
0%100%$40

This means the skill automatically bets more when the edge is larger, and less when the signal is marginal.

版本历史

共 3 个版本

  • v0.0.3 当前
    2026-05-03 03:52 安全 安全
  • v1.0.1
    2026-03-30 01:42 安全
  • v1.0.0
    2026-03-19 19:39

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-intelligence

self-improving agent

pskoett
捕获经验教训、错误和纠正,以实现持续改进。使用时机:(1)命令或操作意外失败;(2)用户纠正……
★ 4,055 📥 795,465
ai-intelligence

ontology

oswalpalash
类型化知识图谱,用于结构化智能体记忆与可组合技能。支持创建/查询实体(人员、项目、任务、事件、文档)及关联...
★ 709 📥 243,478
ai-intelligence

Self-Improving + Proactive Agent

ivangdavila
自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。
★ 1,349 📥 317,642