binance/binance-skills-hub

binance-leaderboard

On-chain wallet leaderboard and Gem Hunter analysis on Binance Web3.

소스 보기

사용 방법

  1. 설치 명령을 복사하고 자신의 에이전트 환경에 설치하기 전에 소스 패키지를 확인하세요.
  2. 필요한 도구, 인증 정보, 지원 환경은 원본 문서에서 확인하세요.
  3. 입력과 원하는 산출물을 설명하고 출처와 결과를 대조하세요.
npx skills add https://github.com/binance/binance-skills-hub --skill binance-leaderboard

시작 전 확인

  • 설치에는 모델 구독, API 한도, 파일 접근 권한이 포함되지 않습니다.
  • 별점 수와 설치 수는 활동 지표이며 독립적인 품질·보안 검증이 아닙니다.
원본 Skill 문서

원본 저장소의 제목, 예시, 코드, 표, 링크, 이미지를 유지해 표시합니다.

Binance Leaderboard Skill

On-chain wallet leaderboard ranking and address analysis. Query top traders by PnL, win rate, and more. Evaluate wallet quality with a 6-dimension scoring model.

Prerequisites

This skill requires the baw CLI (@binance/agentic-wallet npm package). If baw is not found:

bash
npm install -g @binance/agentic-wallet

Verify: baw --version should print 1.6.2 or higher. If installation fails or the user doesn't have Node.js, inform them that Node.js >= 18 is required.

When to Use

User intentCommand
Query top traders by PnL/win rate/volumebaw leaderboard query
Analyze a single wallet address (6-dim score + AI archetype)baw leaderboard analyze
Find wallets holding specific tokens (Gem Hunter)baw leaderboard alpha-radar
Save/load preset filter conditionsbaw leaderboard preset save/list
Save/load Gem Hunter configsbaw leaderboard alpha-radar-config save/list

Supported Chains

ChainchainId
BSC56
SolanaCT_501
Base8453
Ethereum1

Command Tree

baw leaderboard
  query                     # Leaderboard query (Public, no auth)
  analyze                   # Single address analysis (6-dim score + AI overlay)
  alpha-radar              # Gem Hunter query (Private, agentSessionId)
  preset
    save                    # Save preset filters (Private)
    list                    # List preset filters (Private)
  alpha-radar-config
    save                    # Save Gem Hunter config (Private)
    list                    # List Gem Hunter config (Private)

All commands support --json for structured output.

Leaderboard Query

bash
# Basic query — top 20 by PnL on BSC, 7d period
baw leaderboard query -c 56 -p 7d -t ALL --json

# Sort by win rate, KOL tag only
baw leaderboard query -c 56 -p 30d -t KOL --sort-by 20 --json

# Pagination (page from 0, size max 20)
baw leaderboard query -c 56 -p 7d --page 0 --size 20 --json

Public endpoint — no auth required. Returns per-address PnL, win rate, volume, trade count, token distribution, daily PNL, and top earning tokens.

Key query parameters: -c/--chain-id (required), -p/--period (7d/30d/90d), -t/--tag (ALL/KOL/MPC), --sort-by (0=PnL · 20=Win Rate · 30=Total Volume · 50=Trade Count · 60=Recent Activity · 70=Profit Rate · 80=Token Count), --order-by (0/2=Descending · 1=Ascending), --page (from 0), --size (max 20).

Full parameter and return field reference: `references/cli.md`

Single Address Analyze

bash
# Analyze a wallet address — 6-dim scoring + AI archetype
# Default scans top 1000 entries
baw leaderboard analyze -c 56 -a 0xabc... --json

# Scan more entries (up to 5000) for long-tail addresses
baw leaderboard analyze -c 56 -a 0xabc... --top-n 5000 --json

Evaluates the address across 6 dimensions (winrate 25 + stability 20 + drawdown 20 + tags 15 + pnl 10 + followfriendly 10 = 100), then applies an AI overlay (archetype + behaviorflags + ai_adjustment ±10).

Flow: Query leaderboard top 1000 (configurable via --top-n) → reverse-lookup the target address → compute scores → apply AI overlay → output rating.

Rating: ⭐⭐⭐ ≥ 80 · ⭐⭐ ≥ 65 · ⭐ ≥ 50 · ❌ < 50

If the address is not in the top N (default 1000), returns "N beyond top". Use --top-n to increase scan range up to 5000.

Full scoring model details: `references/scoring.md`

Gem Hunter

bash
# Find wallets holding specific tokens
baw leaderboard alpha-radar -c 56 \
  -t 0xtoken1,0xtoken2 \
  -m 1 --json

Private endpoint — requires agentSessionId. Extra required params: -t/--tokens (comma-separated token addresses), -m/--match-count (≥ 1). Supports -p/--period, --page, --size like query.

Returns records with the same fields as leaderboard query, but topEarningTokens replaced by marchedTokens (matched tokens).

Note: The field is spelled marchedTokens (not "matched") in the CLI output.

Preset & Gem Hunter Config

bash
# Save preset filter conditions (pass null/empty to clear all)
baw leaderboard preset save --config '[{"name":"MyPreset","period":"7d","winRateMin":50}]' --json

# List saved presets
baw leaderboard preset list --json

# Save Gem Hunter config (pass null/empty to clear)
baw leaderboard alpha-radar-config save -c 56 \
  --config '[{"uuid":"u1","name":"MyRadar","matchTokenCount":2,"tokenAddressList":[{"tokenAddress":"0xabc"}]}]' --json

# List saved configs
baw leaderboard alpha-radar-config list -c 56 --json

Core Rules

Scoring Model Overview

6-dimension model (total 100) + AI overlay (±10). Dimensions: winrate (25), stability (20), drawdown (20), tags (15), pnl (10), follow_friendly (10). Full tiered scoring tables and AI overlay rules: `references/scoring.md`.

Top-N Reverse Lookup

analyze queries the leaderboard's top 1000 entries (configurable via --top-n, max 5000) and reverse-looks-up the target address. If not found, returns "N beyond top" — inform the user they can increase --top-n or use binance-wallet-tracker's address list for long-tail wallets.

Preset & Config Clearing

Passing null or empty array to preset save or alpha-radar-config save clears all saved items.

Cross-Skill Scenarios

Some scenarios require both leaderboard and wallet-tracker skills:

  • evaluate follow list: Use binance-wallet-tracker's tracker follow to get the user's followed addresses, then leaderboard analyze each one.
  • leaderboard diff: Query leaderboard top N, then diff against binance-wallet-tracker's address list to find untracked wallets.
  • batch import: Query leaderboard with filters → output address list → feed to binance-wallet-tracker's address batch command.

Write-Back Confirmation

After preset/config save operations, re-fetch via list to confirm the operation succeeded — don't assume success from the API response alone. If readback shows the saved data is missing, tell the user "Save may not have taken effect, please try again later" — never mention backend bugs or silent failures.

User-Facing Presentation

This skill serves end users, not developers. Internal field names, error codes, and CLI internals must never appear in user-facing output.

1. Term Mapping (internal → user-facing)

Internal valueUser-facing termNotes
archetype: sniperSniper
archetype: swingSwing Trader
archetype: accumulatorAccumulator
archetype: farmerFarmer
archetype: mixedMixed
sort-by: 0Sort by PnL
sort-by: 20Sort by Win Rate
sort-by: 30Sort by Total Volume
sort-by: 50Sort by Trade Count
sort-by: 60Sort by Recent Activity
sort-by: 70Sort by Profit Rate
sort-by: 80Sort by Token Count
address (in display)omitDon't show raw address unless user asks; use {addressLabel}
finalScoreScoreDon't show the formula totalScore + aiAdjustment

Raw enum values (sniper, swing, etc.) may appear in CLI syntax examples and internal lookup tables, but never in user-facing replies.

2. Never expose internal identifiers

  • groupId, displayOrder — internal IDs, never shown to users
  • Error codes (70001001, etc.) — translate to natural-language messages only
  • Backend behavior details (e.g. preset save returning data: true without persisting) — never mention "backend bug" or "silent failure"

3. Display template hygiene

Use user-facing terms in all output. The archetype field should be translated to its Chinese term (Sniper, Swing Trader, etc.) — never show the raw English enum value. Score should be shown as a number, not as a formula.

Error Codes

Error codes are for internal lookup only — never show numeric codes or internal names to users. Translate to natural-language messages.

CodeInternal NameUser-Facing Message
70001001TRACKERAPIERRORQuery failed, please try again later
70004001TRACKERLEADERBOARDEMPTYLeaderboard data is empty
70004002TRACKERADDRESSNOT_RANKEDAddress not ranked (beyond top 250)

Display Templates

Leaderboard entry:

{addressLabel} | PnL: {realizedPnl} ({realizedPnlPercent}%) | Win Rate: {winRate}% | Trades: {totalTxCnt} | Tokens: {totalTradedTokens}

Address analysis rating:

📊 {addressLabel} Address Analysis

Rating: ⭐⭐⭐ (85/100)
Trading Style: Sniper
Behavior Patterns: High Frequency Small Amount, Nocturnal Active

Dimension Scores:
  Win Rate: 22/25 | Stability: 18/20 | Drawdown: 16/20
  Tags: 15/15 | PnL: 8/10 | Trackability: 6/10

AI Adjustment: +5 (consistent trading style, stable pattern)

Full CLI Reference

관련 작업

같은 저장소의 Skills

더 많은 Skills

모든 Skills
binance
커뮤니티

square-post

Use when the user wants to publish new content to Binance Square — short text, multi-image posts (up to 4), long-form articles with an optional cover, or videos with an auto-generated cover frame. Trigger on direct phrasings like "post to Square", "publish to Binance Square", "发广场", "发布到广场", and on near-miss intents where the user clearly wants to share or publish content on Square even without naming the skill: "share this analysis on Square", "把这篇文章发出去", "发个动态", "把这个视频上传到广场", "publish my chart

24시간 설치 수
364
GitHub Stars
1.2천
업데이트
11월 12일
binance
커뮤니티

binance-agentic-wallet

Use when the user mentions wallet connect/sign in/sign out, check balance, send/transfer tokens, swap/buy/sell tokens, DEX trade, limit/market order, cancel order, get a quote, transaction history, wallet settings, daily limit, slippage, MEV protection, supported chains, send to ENS domain, send to .eth or .bnb name, resolve domain name to address, domain-based transfer, send to .ai Agent ID name, Agent ID claim status, whether I claimed my .ai space id / Agent ID, get my .ai domain, prediction

24시간 설치 수
416
GitHub Stars
1.2천
업데이트
11월 12일
binance
커뮤니티

binance-onchain-copy-trader

Scaffold for assembling on-chain copy-trading strategies on Binance Agentic Wallet. Wires the existing capabilities — address monitoring, signal subscription, address scoring, token audit, order placement, on-chain take-profit/stop-loss — into one runnable pipeline, and exposes every strategy decision as a hook. Ships no strategy of its own. Trigger on: copy trade, copy trading, follow smart money, follow a wallet, mirror trades, copy-trade bot, auto-buy on signal, monitor a wallet and trade, sm

1시간 설치 수
3
GitHub Stars
1.2천
업데이트
11월 12일
binance
커뮤니티

binance-prediction-updown

Scaffold for assembling prediction-market strategies on Binance Agentic Wallet, for recurring BTC Up/Down rounds. Wires round enumeration, market screening, two-step ordering, fill and settlement confirmation and redemption into one runnable pipeline, and exposes every probability judgement as a hook. Ships no strategy of its own. Trigger on: up down prediction, BTC up or down, prediction market bot, hourly prediction rounds, trade up/down automatically, prediction market strategy, prediction ca

1시간 설치 수
3
GitHub Stars
1.2천
업데이트
11월 12일