What problem does it solve?
It helps founders turn creator inputs into on-brand writing by scoring relevance across who-you-are (X), what-you-ship (Y), and your industry edge (Z), then extracting and saving high-signal founder memory before generating output grounded in that memory.
Core Features & Use Cases
- Vector scoring with memory fit: Computes raw axis relevance (0-100), applies memory-fit weighting from existing Markdown memory, and flags weak relevance when magnitude is low.
- Axis-aligned snippet extraction + confirmation gating: Splits submitted content into reusable candidate blocks grouped by X/Y/Z, shows them to the user, and requires explicit confirmation before appending to memory.
- Three-file long-term founder memory management: Maintains x-who-you-are.md, y-product.md, and z-industry-info-gap.md under a configurable memory directory.
- Memory completion & credibility: Calculates log-based completion progress per axis and derives an average credibility score to inform writing quality expectations.
- 3D visualization report: Generates a draggable HTML Canvas report showing the content vector, weak-zone region, and completion bars (run before appending).
- Memory-grounded writing iteration: Drafts founder-brand copy using the three memory files and iterates internally until the output vector magnitude reaches the target threshold.
Quick Start
Use this skill to analyze and prepare memory updates by asking: "评一下这条内容，并提取可保存的X/Y/Z候选片段，生成向量报告后请我确认保存。"
