📦 Part of the offer-toolkit-skill — the full job-hunt bundle (JD · Resume · BQ). Install the bundle to get all three in one shot.
An agent skill for behavioral interview prep. Instead of handing you canned answers, it helps you dig your real past experiences out and organize them into a story bank you can reuse. It asks about your experience step by step, uses STAR/CAR to shape each story, tags them ("took ownership", "handled ambiguity", etc.), and saves each one in English and Chinese — so next time you get a different behavioral question, the same story still works.
JD-driven prep output: a Top 20 question set reverse-engineered from a JD, with STAR templates and editable model answers. (Demo uses a fictional candidate — no real personal data.)
bq-skill/
├── SKILL.md # Entry point: intent routing + five flows
├── prompts/
│ ├── story-mining.md # ★ Four-layer probing engine (story mining)
│ ├── structuring.md # Diagnose + rewrite an existing answer (polishing)
│ └── jd-driven-prep.md # ★ JD × resume → Top 20 questions + STAR templates (JD-driven prep)
├── frameworks/
│ ├── star-car.md # STAR/CAR/SOAR selection + one story, many questions
│ ├── competency-tags.md # Competency dictionary + BQ reverse-lookup
│ └── company-profiles.md # Amazon LP / Meta / Anthropic / OpenAI styles
├── assets/
│ └── bq-prep-report.md # Editorial HTML report spec for JD-driven prep
└── story-bank/ # User asset — one .md per story
├── _index.md # Competency → story reverse-lookup table
├── _story-template.md # Story template
└── convince-team-rewrite.md # Example story (safe to delete)
- Mine a new story — pulls real, quantifiable events out of someone who thinks "I have nothing to talk about," then deepens one into a bank-ready story.
- Answer a specific BQ — check the bank first, reuse a hit, or mine a new one on the spot.
- Polish an existing answer — diagnose structure problems (Situation too long, "I" vs "we", no quantified result) and rewrite.
- Mock interview — one question at a time in the target company's style, with feedback.
- JD-driven prep — reads a JD + your resume, reverse-engineers a Top 20 interview-question set for that role, and builds a STAR prep template for each question. (See next.)
JD-driven prep reuses the Must-Haves + Hidden Signals decoded by job-description-skill to reverse-engineer a Top 20 interview-question set for one specific role, builds a STAR prep template per question from your real resume, and outputs an HTML report.
Tell Claude "help me prep for behavioral interviews", "mine a story for my bank",
"here's a BQ, help me answer it", or just run /bq-skill.
- Bank first. Always check
story-bank/_index.mdbefore mining — reuse beats re-digging. - One question at a time. Mining is a conversation, not a questionnaire.
- Never fabricate. All material comes from the user's real experience; every metric is verified with them.
The probing scripts in prompts/story-mining.md and the competency dictionary in
frameworks/competency-tags.md are where to inject your own phrasing and vocabulary — the more you
put in, the more the skill sounds like you.
- offer-toolkit-skill — the all-in-one bundle (JD · Resume · BQ)
- job-description-skill — Job Description Decoder
- resume-builder-skill — Resume generation & beautification
MIT — fork it, remix it, ship your own version.
Created by Dreameryanyan · LinkedIn · X · Xiaohongshu