story-bank: One Project, Many Interview Stories

Where this fits

A candidate has one flagship project but is interviewing for three different roles — AI product, growth, and investment. Each interview wants a different angle, but the numbers must stay consistent. story-bank solves this by mining raw experience into a single fact-based story bank, then repackaging the same facts per role type. It triggers on phrases like 整理项目经历 / 挖掘亮点 / 经历库, and steps in whenever interview-script lacks raw material to write from.

Why agents benefit

• Elicits deeper material: guiding questions (e.g., "what was the messiest state when you took over?" or "what did you decide that no one else would have?") pull out decisions, numbers, and contrast — not just job descriptions.
• Enforces sourced numbers: every figure must trace to a report or retrospective, so the agent never fabricates metrics.
• Repackages per role: the same project card renders different angles for AI product, growth, or investment roles while keeping one fact base.
• Builds a reusable artifact: the persisted story-bank.md feeds later interview-script work, so nothing is re-mined twice.
• Avoids re-asking: the agent reads profile.md and existing story-bank.md first, then only asks what it doesn't know.

Practical scenarios

Interview script has no material

The user asks "这段经历怎么讲" and interview-script has nothing to write from. story-bank runs first, mines the experience into cards, then hands the structured cards back to interview-script for final wording.

Same project, three roles

The user is applying to AI product, growth, and investment roles with one project. story-bank produces one card with the same numbers, then three repackaged versions — model selection and effect validation for product, funnel and ROI for growth, market judgment for investment.

Organizing years of scattered experience

The user says "整理项目经历" and has notes everywhere. story-bank walks through each experience with 2-3 questions at a time, applies the "decision + number + contrast" bar, and merges or drops experiences that don't qualify.

Add it to your agent workflow

npx skills add YunyueLi/greenroom --skill story-bank

Trigger it when the user says 整理项目经历, 挖掘亮点, or 同一个项目怎么对不同岗位讲. The skill reads profile.md and intel.md first, runs the elicitation loop, and writes structured cards to story-bank.md.

Compare before adopting

Compare against generic interview-prep prompts. Check whether the skill actually enforces sourced numbers and role-based repackaging, or just collects anecdotes. Review the GitHub repo for maintenance freshness and whether the elicitation question set fits your workflow. Alternatives like hand-written story templates lack the "fact base once, packaging varies" discipline that makes this skill distinctive.

Why it is worth tracking

At 11 stars, story-bank is early but opinionated: the fact-base-once principle is exactly what interview-prep agents miss. Watch for community adoption of its elicitation questions and whether the markdown card format gains traction across other skills. Evaluate it when your interview-script work keeps stalling on missing raw material. More context at https://www.openagentskill.com/skills/yunyueli-story-bank.