Stablecoin Depeg Scanner 🔍
3-minute crisis arbitrage evaluator. When a stablecoin loses its peg, run the assessment to decide: buy, watch, or avoid.
Philosophy
Depeg = panic pricing. The question is NOT "is it safe" but "is the discount bigger than the real damage?" Small position, high odds, asymmetric payoff.
Quick Start
When user mentions a stablecoin depeg event:
Step 1: Run the automated evaluator (60 seconds)
CODEBLOCK0
- -
<COIN>: CoinGecko ID or symbol (e.g., resolv-usr, USR, DAI) - INLINECODE4 : User's total available capital in USD (default: 5000)
- INLINECODE5 : Target peg price if not $1.00 (default: 1.0)
The script auto-fetches: current price, market cap, TVL from DefiLlama, and calculates odds + position size.
Replace SKILL_DIR with the resolved absolute path of this SKILL.md's parent directory.
Step 2: Complete exploit analysis (60 seconds)
The script outputs a checklist. Use web_search to answer these questions:
- 1. Was collateral stolen or was it a mint/logic exploit?
- Search:
"<coin_name> exploit PeckShield" or
"<coin_name> hack SlowMist"
- Mint/logic exploit (collateral intact) → bullish
- Collateral drained → bearish
- 2. Has the team responded?
- Check their Twitter for pause announcement
- Response within 1h → good sign
- Silent >6h → red flag
- 3. Is the bug fixable?
- Smart contract bug with clear fix → recovery likely
- Fundamental design flaw → avoid
Step 3: Adjust and deliver recommendation (60 seconds)
Based on exploit analysis, adjust the script's score:
- - Collateral confirmed intact → score +10
- Team responded fast with plan → score +10
- Collateral partially stolen → score -10
- No team response → score -20
- Algorithmic stablecoin / no real backing → score -50 (AVOID regardless)
Deliver final recommendation with:
- - Action: BUY / WATCH / AVOID
- Position size in USD
- Entry price range
- Exit target
- Stop condition
Decision Framework
For detailed scoring logic, exploit type analysis, historical benchmarks, and red/green flags, read references/decision-framework.md.
Output Format (to user)
CODEBLOCK1
Key Rules
- 1. Speed over perfection — 3 minutes, not 30 minutes. First-mover advantage is everything.
- TVL is truth — DefiLlama TVL > official statements. Contracts don't lie.
- Small positions only — Never >6% of capital on a single depeg bet.
- Algorithmic = death — No real collateral = no recovery. Luna/UST is the permanent lesson.
- Exit plan before entry — Know your sell condition before buying.
稳定币脱锚扫描器 🔍
3分钟危机套利评估器。当稳定币脱锚时，运行评估以决定：买入、观望或回避。
核心理念
脱锚 = 恐慌定价。关键问题不是是否安全，而是折扣是否大于实际损失？小仓位、高赔率、非对称收益。
快速启动
当用户提及稳定币脱锚事件时：
步骤1：运行自动评估器（60秒）
bash
cd SKILLDIR && python scripts/depegeval.py --capital
- - ：CoinGecko ID或符号（例如：resolv-usr、USR、DAI）
- --capital：用户可用总资金（美元，默认：5000）
- --peg：目标锚定价格（若非1.00美元，默认：1.0）
脚本自动获取：当前价格、市值、DefiLlama的TVL，并计算赔率及仓位大小。
将SKILL_DIR替换为此SKILL.md所在目录的解析绝对路径。
步骤2：完成漏洞分析（60秒）
脚本输出一个检查清单。使用web_search回答以下问题：
- 1. 抵押物被盗还是铸造/逻辑漏洞？
- 搜索：
name> exploit PeckShield 或 name> hack SlowMist
- 铸造/逻辑漏洞（抵押物完好）→ 看涨
- 抵押物被掏空 → 看跌
- 2. 团队是否已回应？
- 查看其Twitter是否有暂停公告
- 1小时内回应 → 好信号
- 超过6小时无回应 → 危险信号
- 3. 漏洞是否可修复？
- 智能合约漏洞且有明确修复方案 → 可能恢复
- 根本性设计缺陷 → 回避
步骤3：调整并给出建议（60秒）
基于漏洞分析，调整脚本评分：
- - 抵押物确认完好 → 评分+10
- 团队快速回应并有计划 → 评分+10
- 抵押物部分被盗 → 评分-10
- 团队无回应 → 评分-20
- 算法稳定币/无真实抵押 → 评分-50（无论如何都回避）
给出最终建议，包含：
- - 操作：买入 / 观望 / 回避
- 仓位大小（美元）
- 入场价格区间
- 退出目标
- 止损条件
决策框架
有关详细评分逻辑、漏洞类型分析、历史基准及红/绿信号，请阅读references/decision-framework.md。
输出格式（面向用户）
🔍 稳定币脱锚评估:
📊 价格:
• 当前: $X.XX | 锚定: $1.00 | 偏离: XX%
• 赔率: X.Xx
🏦 抵押物:
• TVL: $XXM (较昨日 -X%)
• 状态: 完整/部分损失/已掏空
🔐 攻击分析:
• 类型: 铸造漏洞/预言机/抵押物被盗
• 团队响应: 已暂停/未回应
• 可修复性: 高/中/低
📋 结论:
• 操作: 🟢买入 / 🟡观望 / 🔴回避
• 仓位: $XXX (总资金X%)
• 目标: $X.XX
• 止损条件: XXX
关键规则
- 1. 速度优先于完美 — 3分钟，而非30分钟。先发优势至关重要。
- TVL即真相 — DefiLlama的TVL优于官方声明。合约不会撒谎。
- 仅限小仓位 — 单次脱锚押注不超过总资金的6%。
- 算法币=死亡 — 无真实抵押=无法恢复。Luna/UST是永久教训。
- 入场前制定退出计划 — 买入前明确卖出条件。
