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 DuanZGit/D-Hub --skill gushengit clone --depth 1 https://github.com/DuanZGit/D-HubWrote 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/duanzgit/d-hub/gushen)<a href="https://agentmods.dev/skills/duanzgit/d-hub/gushen"><img src="https://agentmods.dev/badge/skills/duanzgit/d-hub/gushen/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/duanzgit/d-hub/gushen"><img src="https://agentmods.dev/badge/skills/duanzgit/d-hub/gushen.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 | $0.00110 | $0.04671 |
| Opus 5.5 | $0.00044 | $0.01868 |
| Sonnet 5.5 | $0.00022 | $0.00934 |
| Haiku 4.5 | $0.00011 | $0.00467 |
Grade A, and why
gushen 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 6d 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 — 245 lines — stays where its author put it; the contents beside it link to each section on GitHub.
路径约定：本文件里
$SKILLS= 技能根目录。执行前先解析：SKILLS="$HOME/.agents/skills"; [ -d "$SKILLS" ] || SKILLS=/var/minis/skills（UG/本机命中~/.agents/skills，Minis 命中/var/minis/skills。）
触发条件（硬边界）
只有用户明确提到 gushen 或 股神 时，才激活 gushen。
- 触发词：
gushen、股神
不触发 gushen 的情况：
- 普通搜索/查资料 → 走 agent-reach
- 食谱/健康/生活技巧 → 走 DK
- 深度调研 → 走 research
- 即使用户问的是股票/行情/基金，只要没提
gushen/股神，就不触发 gushen
执行协议（唯一权威）
完整 /gushen 必须先读取并遵守：references/execution-contract.md（Gate A-E）。
本文件的流程说明不得与契约冲突，冲突以契约为准。
硬门槛： 每次完整分析必须出现且仅出现 8 个 Agent（统一 5 分制），缺一或分制错误即未完成。
| # | 角色 | 对齐上游 | 纪律职责（必出 1 句） |
|---|---|---|---|
| 1 | 巴菲特 | financial-analyst | 财务+估值，必市值验算 |
| 2 | 段永平 | business-analyst | 商业模式 + 镜子测试 |
| 3 | 芒格 | industry-researcher | 护城河 + 逆向检验 |
| 4 | 李录 | risk-assessor | 长期确定性 + 10年终局 |
| 5 | 技术分析师 | — | 趋势+支撑阻力 |
| 6 | 动量交易员 | — | 量价+突破 + 反共识 |
| 7 | 情绪分析师 | — | 资金流+拥挤度 + 反共识 |
| 8 | 风控官 | — | 仓位+Kelly符号+否决权 |
纪律（效能来源，ai-berkshire）： 每 Agent 末句 1 句逆向/反共识（价值组"什么情况下我错"，Trading 组"聪明人为何反向"）；灰色地带是合法结论，不硬给价位。
详见：references/execution-contract.md。
Price is what you pay, value is what you get. — Warren Buffett
报告视觉规范（唯一风格）
所有 HTML 报告必须直接套用 templates/daily_report.html v4 lean，不得回退旧版或自创设计。
- 设计定位：移动金融研究简报，
DESIGN_VARIANCE 5 / MOTION_INTENSITY 5 / VISUAL_DENSITY 6。 - 主题：暗色单主题 bg
#0b0e13/ surface#12161f/ surface-2#171c27。 - 强调色：琥珀金
#e8b04b单一，禁紫蓝渐变。 - 涨跌色：红
#ff4d5e=涨/盈利；绿#00c767=跌/亏损，全页锁死。 - 禁 emoji：Agent 用姓氏方块，禁图标。
- 数字：
.numtabular-nums，字阶 11/13/15/26px，圆角 10px。 - 信息架构（精华优先）： 现在做什么 → 短中期三窗口 → 现在发生什么（指数/组合/板块/事件/技术+freshness）→ 为什么（价值汇总→Trading 汇总→8 Agent 明细→交叉验证）→ 最后复盘（历史→风险/质量→总结）。
- 固定顺序：masthead → gates A-E → 今日行动 → 短中期 → 市场状态 → 价值汇总 → Trading 汇总 → 8 Agent → 交叉验证 → 历史复盘 → 风险/质量 → 最终总结 → footer。
- Three.js：可选低功耗，必须有文字降级，不承载唯一事实。
- 外部依赖：零 CDN，模板改动过
validate_report.py。
设计原则
- 数据用工作流：
workflow.py单入口（幂等+重试+旧文件兜底），growth/pe自动算不手填，精算前置直供 Agent - 决策用纪律：脚本不替决策，纪律（否决/镜子/逆向/反共识）由提示词+校验器强制。
- 独立判断不靠并行：价值组禁 MA/RSI/量比/MACD/VIX，Trading 组禁 PE/ROE/EPS/股息，校验器字符级拦。
- 精华优先：操作建议前置，纪律细节折叠，报告总长不因纪律而膨胀。
- 红涨绿跌 / 次日作废 / 移动端优先 / 先指数后个股 / 持仓必结合（保持 v2）。
What ships with it
23 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- assets/three.module.min.js 331 KB runs code
- config/finnhub.yaml 397 B
- DATA_SOURCES.md 1.9 KB
- references/event-calendar.md 2.6 KB
- references/execution-contract.md 3.0 KB
- references/report-template.md 1.9 KB
- scripts/load_key.sh 1.3 KB runs code
- skills/technical-analysis.md 3.1 KB
- templates/daily_report.html 31 KB
- templates/gushen_report_20260901.html 42 KB
- templates/gushen_report_20260902.html 41 KB
- templates/gushen_report_20260903.html 46 KB
- tools/adanos.py 9.8 KB runs code
- tools/backtest.py 5.7 KB runs code
- tools/cn_market.py 32 KB runs code
- tools/deep.py 8.3 KB runs code
- tools/dump_facts.py 4.0 KB runs code
- tools/financial_rigor.py 18 KB runs code
- tools/technical_indicators.py 10 KB runs code
- tools/validate_contract.py 582 B runs code
- tools/validate_report.py 2.9 KB runs code
- tools/verify.py 9.2 KB runs code
- tools/workflow.py 14 KB runs code
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.
- 6d ago First seen · 245 lines · 110 tokens per session scan A 9c22726d03ad
gushen is a skill published in the GitHub repository DuanZGit/D-Hub (0 stars, last pushed today), licensed MIT. It adds 110 tokens to every session and 4,671 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-11-06.
gushen compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| gushenDuanZGit/D-Hub | Skill | 0 | today | MIT | A | 110 | Claude Code, Codex |
| tradingagents-analysisKylinMountain/TradingAgents-AShare | Skill | 899 | 14d ago | none | A | 138 | Claude Code, Codex |
| comps-analysisginlix-ai/LangAlpha | Skill | 2.3k | today | Apache-2.0 | A | 48 | Claude Code, Codex |
| dcf-modelginlix-ai/LangAlpha | Skill | 2.3k | today | Apache-2.0 | A | 56 | Claude Code, Codex |
| company-profileginlix-ai/LangAlpha | Skill | 2.3k | today | Apache-2.0 | A | 68 | Claude Code, Codex |
| check-deckginlix-ai/LangAlpha | Skill | 2.3k | today | Apache-2.0 | A | 63 | Claude Code, Codex |
| check-modelginlix-ai/LangAlpha | Skill | 2.3k | today | Apache-2.0 | A | 72 | Claude Code, Codex |