Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Lzh-xbccz/hermes-finance --skill microcap-pnd-systemgit clone --depth 1 https://github.com/Lzh-xbccz/hermes-financeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/lzh-xbccz/hermes-finance/microcap-pnd-system)<a href="https://agentmods.dev/skills/lzh-xbccz/hermes-finance/microcap-pnd-system"><img src="https://agentmods.dev/badge/skills/lzh-xbccz/hermes-finance/microcap-pnd-system/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/lzh-xbccz/hermes-finance/microcap-pnd-system"><img src="https://agentmods.dev/badge/skills/lzh-xbccz/hermes-finance/microcap-pnd-system.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00071 | $0.03399 |
| Opus 5 | $0.00036 | $0.01699 |
| Sonnet 5 | $0.00014 | $0.00680 |
| Haiku 4.5 | $0.00007 | $0.00340 |
Grade A, and why
microcap-pnd-system scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 227 lines — stays where its author put it; the contents beside it link to each section on GitHub.
极端山寨P&D交易系统 v3.1
⛔ 铁律：你不是在写行情播报，你是在做证据审计。每个方向判断必须回答"为什么"，并先做反向审计。允许使用"证据不足"、"方向未确认"、"观望"；禁止在证据冲突时硬给做多/做空。
核心认知：86%的极端山寨最终崩盘>50%。赚钱靠的不是选对币，是在崩之前跑掉。 铁则：每次分析必须全维度，禁止只给价格就建议入场。泵幅>50%自动降级。
一、系统架构
pnd_detect.py v3（全自动深度分析）
├── ① Binance现货 24hr ticker
├── ② 4H/1H/日线 K线 + 量轨迹
├── ③ 永续合约 OI + 资金费率(5次历史)
├── ④ CoinGecko: MCAP/FDV/流通/ATH/ATL/排名
├── ⑤ 社区: Twitter粉丝数 + 项目分类
├── ⑥ 退市风险检查 (已知退市列表 + 描述关键词)
├── ⑦ 垃圾币自动过滤 (ATH-99% + 死社区 + 僵尸量 + 排名>1500)
├── ⑧ 红牌检查 (换手/流通/FDV/MCAP/ATH回撤)
├── ⑨ 量价健康度诊断 (量缩率 + 上影 + 换手率)
└── ⑩ 综合评分(0-10) + 交易建议
pnd_track.py v3（全维度实时追踪）
└── 价格/量/OI/费率/阶段变迁/健康度/预警 (Binance→Bybit→Gate.io回退)
pnd_exit.py v3（全维度退出决策引擎）
└── 5级信号权重 + 6级操作指令 + OI背离检测 + 移动止损
pnd_whale.py v1（BSC链上鲸鱼监控）
└── 持仓集中度 + 大额转账 + 鲸鱼动向检测
二、脚本详解
2.1 怎么抓 — pnd_detect.py v3.0（全自动深度分析）
一条命令 = 扫描 + 基本面 + 合约 + 垃圾币过滤 + 评分 + 建议。
# 全自动扫描+深度分析Top8
python3 ~/.hermes/scripts/microcap/pnd_detect.py
# 指定币种深度分析
python3 ~/.hermes/scripts/microcap/pnd_detect.py --symbol BANANAS31,EDEN,SPELL
# 只看能进的(评分≥5)
python3 ~/.hermes/scripts/microcap/pnd_detect.py --deep 10 --min-score 5
# JSON输出
python3 ~/.hermes/scripts/microcap/pnd_detect.py --json --deep 5
# 后台定时扫+预警文件输出
python3 ~/.hermes/scripts/microcap/pnd_detect.py --deep 5 --alert
评分体系：
- ≥7.0 → 🟢 可进（信号强，风险可控）
- 5.0-6.9 → 🟡 谨慎（有信号但存在风险点）
- 3.0-4.9 → 🟠 高风险（多项红牌，小仓位可赌）
- 1.0-2.9 → 🔴 不建议（风险大于机会）
- <1.0 → 💀 不做（垃圾币/退市候选）
垃圾币自动过滤器（任一满足即扣分至负）：
- ATH回撤 > 99.5%（死猫）
- CoinGecko排名 > 1500（无人关注）
- Twitter < 1K粉（死社区）
- 24h换手率 < 0.01（僵尸）
- 已知退市记录（Bitget/Bitfinex等）
9级P&D阶段判定：
| 阶段 | 条件 | 操作 |
|---|---|---|
| 🟢 EARLY | 24h涨10-15%，高回撤<3%，量↗ | ✅可进 |
| 📈 PUMP | 24h涨15-40%，高回撤<8%，量↗ | ✅可追 |
| 🚀 SURGE | 1h涨>15%，量↗↗ | 🟢可追 |
| 🚀 MOON | 24h涨>40%，高回撤<8% | 🟢可追 |
| 🛑 ACCUM | 横盘，换手0.02-0.2x | 🟡观察 |
| ⚠️ FADE | 高回撤25-40% | 🔴回避 |
| 💀 DIST | 高回撤>40% | 🔴回避 |
| 🔻 DUMP | 24h跌>20% | 🔴回避 |
| 😴 ZOMBIE | 换手<0.01 | 💀不做 |
2.2 怎么跟 — pnd_track.py v3.0（全维度追踪）
新增：OI+价格四象限、健康度评分(0-100)、阶段变迁标注、入场信号检测
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 227 lines · 71 tokens per session scan A 974019da6302
microcap-pnd-system is a skill published in the GitHub repository Lzh-xbccz/hermes-finance (24 stars, last pushed 2mo ago), licensed MIT. It adds 71 tokens to every session and 3,399 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
copy-trader
Social trading - copy the best traders automatically. Track whales, influencers, and top performers.
defi-connector
Connect to DeFi protocols for staking, lending, yield farming, and liquidity provision. Full auto-compound support.
defi-yield
DeFi yield hunting across protocols. Find the best APY, auto-compound, manage LP positions, and optimize gas.
defi-dashboard
Skill #62 | Category: DeFi | Inspired by: Zapper, DeBank, Zerion.
signal-copier
Copy trading signals from Telegram channels and execute them automatically across ALL supported markets.
social-feed
Skill #63 | Category: Social | Inspired by: Zapper Feed, eToro Social, Nansen Smart Money.