菜单
Information-transfer arbitrage for prediction markets — detect when market price deviates from superior reference price and capture convergence. Not forecasting.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Farrice/mes-knowledge-base --skill prediction-market-ai-event-analysis命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
The /generate front door — in-house, pay-as-you-go creative generation (Higgsfield replacement). One command for image, video, and audio: routes each request to the right model recipe (skills/generate/models/*.json), defers to binding creative_router lanes (people → fal-people, style-family → fantastic-posters; Higgsfield fully retired 2026-08-06), quotes paid video before running, honors prompt-level budgets ("total budget $3"), runs comparison batches across models, logs every prompt beside its asset, and auto-refreshes the Asset Command Center (/assets-board). Use when the user says "generate", "make an image/video/voiceover", "create ads/assets/creative", names a model (recraft, kling, seedance, nano banana, gpt image), or asks for a multi-model comparison. Source: RoboNuggets 9C4TRbucmhQ, rebuilt on our fal-first stack.
The Edit Bay — in-house conversational video studio (VOX-style explainers, documentaries, shorts, LinkedIn video). Transcript-driven cutting, B-roll ladder, motion graphics as code, captions, music, publish-ready packages — all agent-operated, free-first. Use when Farrice says "edit this video/footage", "make an explainer/documentary/short", "cut this recording", "add B-roll/captions/music", "turn this essay/post into a video", or any video production ask. Front door for three modes: talking-head, VO-only, zero-camera (his recorded VO + assembled visuals — NO voice clone, NO TTS narration, his ruling 2026-08-06). Source: Brad Bonanno agentic edit pipeline (extractions/brad-bonanno-edit-bay/), rebuilt on the house stack: WhisperX local + edit_bay.py ffmpeg assembly + HyperFrames/Remotion graphics + /generate (fal) — NO Higgsfield (retired 2026-08-06).
Architects technical explainer videos using Brad Bonanno's structural patterns — survives short attention spans, communicates complete systems (not just teasers), and functions as a channel-building artifact (one video pulls subscribers, not just views). Use when planning a long-form technical explainer (5-20 min) for YouTube, structuring an explanation of a complex system the user knows but can't teach yet, balancing depth vs accessibility for a technical audience, or building an evergreen video meant to compound subscriber growth over years rather than spike then die. Trigger proactively whenever the user says "explainer video", "technical explainer", "how do I explain X", "long-form YouTube", "evergreen video", or has a complex system that needs to land with a non-expert. For Instagram-format shareworthy content use brock-johnson-shareworthy-content; for short-form video patterns use a TikTok/Reels-specific skill.
基于 SOC 职业分类
正在显示 SKILL.md
| name | Prediction Market AI Event Analysis |
| description | Information-transfer arbitrage for prediction markets — detect when market price deviates from superior reference price and capture convergence. Not forecasting. |
| version | 1.0 |
| format | completion-engine |
| workflows | 3 |
| source | MES 3.0 Deep Extraction — sovereign2013 + PolySwarm + ensemble bots (11 sources, 607 lines) |
Prediction market trading is NOT a forecasting problem. It is an information-transfer arbitrage problem. The 7.6% of wallets that profit do not predict outcomes better than the market. They detect when the market's price deviates from a superior reference price (sportsbook odds, exchange spot prices, multi-model ensemble consensus) and capture the convergence.
sovereign2013's $1-to-$3.3M run is the proof case: 37,247 bets in ~8 months, almost exclusively sports, Claude-powered. The bot does not "know" sports. It knows that Vegas knows sports, and that Polymarket participants don't know what Vegas knows.
Three workflows covering the full prediction market trading pipeline:
Odds Discrepancy Scanner (workflows/odds-discrepancy-scanner.md) — Scans active Polymarket sports/event markets against sportsbook reference odds. Converts to implied probabilities, identifies gaps exceeding fee drag, outputs ranked opportunity list with confidence levels and action recommendations.
Edge Validation & Sizing (workflows/edge-validation-sizing.md) — Takes an identified discrepancy through a 6-point validation framework: data integrity, gap explanation, fee-adjusted edge, position sizing (quarter-Kelly), timing check, and kill conditions. Outputs a validated trade plan or rejection with reasons.
Multi-Model Ensemble (workflows/multi-model-ensemble.md) — Runs independent analysis through 3 weighted perspectives (analytical 40%, credibility 35%, contrarian 25%), aggregates via Bayesian mixture with 30% market weight, outputs probability estimate with confidence, disagreement analysis, and trade recommendation.
| Fact | Number | Source |
|---|---|---|
| Profitable wallets | 7.6% of 50,000+ | Wallet analysis |
| Bot dominance | 14 of 20 top wallets | 2026 data |
| Arbitrage window (2024) | 12.3 seconds avg | Market data |
| Arbitrage window (2026) | 2.7 seconds avg | Market data |
| Paper-to-live gap | 522x paper vs -49.5% live | Jung-Hua Liu v2 |
| Annual extraction | ~$40M from Polymarket | Apr 2024-Apr 2025 |
| Human underperformance | ~18% vs bots on same strategy | Behavioral analysis |
The paper-to-live gap is the central risk. Multiply any paper edge by 0.5-0.7 before committing capital. If it doesn't survive the haircut, it isn't real.
Execution is 70% of success. Slippage (2-4 cents), fees (0.75-2%), latency, and liquidity collectively destroy edges that look massive on paper.
Start with sports arbitrage, not crypto. Sports have the best reference prices (sportsbook odds), the widest Polymarket inefficiency, and don't require sub-100ms infrastructure.
92.4% of wallets are unprofitable. The three killers: oversized positions, late entries, inconsistent risk management.
| Model | Weight | Strength |
|---|---|---|
| GPT-4o | 40% | Broad analytical reasoning, statistical pattern recognition |
| Claude | 35% | Source credibility evaluation, nuanced uncertainty reasoning |
| Gemini | 25% | Contrarian perspective, alternative data interpretation |
Edge = (Your_Probability - Market_Price) / (1 - Market_Price)
Position = 0.25 * Edge * Bankroll
Cap at min(Position, 0.05 * Bankroll)
Hard caps: 5% single position, 15% correlated exposure, 5% daily drawdown circuit breaker.
| Segment | Speed Required | Reference Price | Who Wins |
|---|---|---|---|
| Ultra-short crypto (5-15 min) | Sub-100ms | Exchange spot | Infrastructure bots only |
| Sports events | Seconds | Sportsbook odds | Claude-speed bots |
| Political/economic | Minutes | Polling/models | AI ensembles + judgment |
genius.md contains: 12 Genius Patterns, 10 Hidden Knowledge items, 3 Exemplars + 1 Anti-Exemplar, 7 Signature Moves, and the Quality Rubric. Load for deep analysis, complex market evaluation, or when output quality needs to exceed score 7.
3 deterministic practitioner prompts — each carries an Output Contract, Output Skeleton, and Quality Gate. When a deliverable matches one, Read it and honor its contract instead of improvising the output shape.
skills/prediction-market-ai-event-analysis/references/prompts-v2/edge-validation-trade-plan.mdskills/prediction-market-ai-event-analysis/references/prompts-v2/multi-model-ensemble-forecast.mdskills/prediction-market-ai-event-analysis/references/prompts-v2/odds-discrepancy-scan.md