Skip to content

AIPMAndy/BPClaw

Repository files navigation

🎯 BPClaw

Research-Driven Fundraising BP Document Agent

Transform "writing a BP" from a one-time task into a reusable workflow with scoring loops and multi-format output.

License GitHub Stars Python

English | 简体中文


🚀 Why BPClaw?

Capability Traditional BP Writing BPClaw
Research Evidence Fragmented Systematic (macro/micro + source logging)
Structure Completeness Depends on experience Fixed workflow (6 chapters + scoring criteria)
Actionability Heavy narrative, weak execution Forced milestones, owners, stop-loss conditions
Reusability Hard to accumulate Scripted templates, continuous iteration

⚡ 30-Second Quick Start

# 1) Initialize BP workspace
python3 skills/public/bpclaw/scripts/init_bp_project.py "my-startup-bp" --out .

# 2) Score your BP draft
python3 skills/public/bpclaw/scripts/score_bp.py \
  --input my-startup-bp/outputs/2_bp_draft.md \
  --output my-startup-bp/outputs/4_bp_review_scorecard.md

# 3) Generate LivePPT scene map from BP draft
python3 skills/public/bpclaw/scripts/generate_liveppt_scene_map.py \
  --input my-startup-bp/outputs/2_bp_draft.md \
  --output my-startup-bp/outputs/3_liveppt_scene_map.md

# 4) Optional: Export to DOCX (requires pandoc)
python3 skills/public/bpclaw/scripts/export_doc.py \
  --input my-startup-bp/outputs/2_bp_draft.md \
  --output my-startup-bp/outputs/2_bp_draft.docx

📋 Workflow

  1. Input background and funding goals
  2. Deep macro/micro research with evidence chain logging
  3. First-principles analysis (Musk Drill)
  4. Output BP draft with auto-scoring
  5. Export DOCX and LivePPT scene map
  6. Review and accumulate for next iteration

🤝 Author

Andy | AI Product Expert

  • 🚀 Ex-Tencent / Ex-Baidu AI Product Lead
  • 🦄 LLM Unicorn VP → Startup CEO
  • 🎯 AI Business Strategy Consultant

WeChat: AIPMAndy | GitHub: @AIPMAndy


🎯 BPClaw

深度调研驱动的融资 BP 文档 Agent 体系。
把"写 BP"从一次性文档任务升级为"可复用流程 + 评分闭环 + 多格式产出"。

为什么需要 BPClaw

能力 传统 BP 写作 BPClaw
调研证据链 片段化 系统化(宏观/微观 + 来源日志)
结构完整度 依赖个人经验 固定流程(6 大章节 + 评分标准)
可执行性 叙事多、落地弱 强制里程碑、owner、止损条件
二次复用 难以沉淀 脚本化模板,可持续迭代

30 秒快速开始

# 1) 初始化 BP 工作区
python3 skills/public/bpclaw/scripts/init_bp_project.py "textile-financing-bp" --out .

# 2) 在 outputs/2_bp_draft.md 填入内容后评分
python3 skills/public/bpclaw/scripts/score_bp.py \
  --input textile-financing-bp/outputs/2_bp_draft.md \
  --output textile-financing-bp/outputs/4_bp_review_scorecard.md

# 3) 从 BP 草案自动生成 LivePPT 场景图
python3 skills/public/bpclaw/scripts/generate_liveppt_scene_map.py \
  --input textile-financing-bp/outputs/2_bp_draft.md \
  --output textile-financing-bp/outputs/3_liveppt_scene_map.md

# 4) 可选:导出 DOCX(需要本机安装 pandoc)
python3 skills/public/bpclaw/scripts/export_doc.py \
  --input textile-financing-bp/outputs/2_bp_draft.md \
  --output textile-financing-bp/outputs/2_bp_draft.docx

核心命令

python3 skills/public/bpclaw/scripts/init_bp_project.py "<project-name>" --out .
python3 skills/public/bpclaw/scripts/score_bp.py --input <bp.md> --output <scorecard.md>
python3 skills/public/bpclaw/scripts/generate_liveppt_scene_map.py --input <bp.md> --output <scene-map.md>
python3 skills/public/bpclaw/scripts/export_doc.py --input <bp.md> --output <bp.docx>

工作流

  1. 输入背景与融资目标。
  2. 宏观/微观深调研并记录证据链。
  3. 第一性原理拆解(Musk Drill)。
  4. 输出 BP 草案并自动评分。
  5. 导出 DOCX 与 LivePPT 场景图。
  6. 复盘沉淀为下一轮自动化能力。

Roadmap

  • v0.1.x:完成 BP 工作流与评分脚本首版。
  • v0.2.x:补齐开源发布文档与 CI smoke 校验。
  • v0.3.x:加入行业模板包(制造业/消费品/企业服务)。
  • v0.4.x:加入指标抽取与自动证据完整性检查。

贡献

License

Apache-2.0 + ADDITIONAL TERMS / 附加条款。

作者

AI酋长Andy | 前腾讯/百度 AI 产品专家

  • 🚀 大模型独角兽 VP → 创业 CEO
  • 🎯 AI 商业战略顾问

微信: AIPMAndy | GitHub: @AIPMAndy


⭐ If this helps, please give it a star! / 觉得有用请点个 Star!

About

🎯 融资BP文档Agent体系 — 深度调研驱动、结构化评分、多格式产出

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages