# Interview Prep

> Use when preparing for a specific interview opportunity, round, interviewer, panel, case, project deep dive, technical discussion, behavioral interview, or hiring-manager conversation.

- Skill: `zhanlincui/interview-prep` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add zhanlincui/interview-prep`
- Raw SKILL.md: https://api.skillmd.com/api/skills/zhanlincui/interview-prep/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: ZhanlinCui (https://skillmd.com/u/zhanlincui)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/zhanlincui/interview-prep

---


# Interview Preparation

**REQUIRED BACKGROUND:** Use `career-state-protocol`. Use the current `T###` Target Model from `jd-company-research` and stories from `story-bank-builder`.

## Goal

Build a stage-specific preparation pack that aligns what the target is likely to test with what the candidate can prove and say naturally.

## Mandatory inputs

Collect these before a final prep pack:

- company, role, and actual JD or Target Model;
- candidate Career Profile, project inventory, or resume;
- submitted resume version when one exists;
- interview stage, format, date, and language;
- known interviewer roles and prior-round feedback when available.

Ask the highest-value missing item one question at a time. Under time pressure, produce a clearly provisional priority brief while collecting the remaining inputs.

## Workflow

1. Refresh time-sensitive target facts if the existing research may be stale.
2. Read the `T###` capability model and infer the purpose of this stage.
3. Run consistency checks between the submitted resume and Career State.
4. Build a question map from target evidence, candidate gaps, resume claims, prior feedback, interviewer function, and stage.
5. Build a story map with primary story, backup story, claim support, and follow-up risk.
6. Route weak project evidence to `project-deep-dive` and missing coverage to `story-bank-builder`.
7. Prepare answer theses, not invented scripts, for questions with supported evidence.
8. Add skeptical follow-ups for ownership, attribution, tradeoffs, AI depth, failure, adoption, and scale as relevant.
9. Create a training plan ordered by expected interview impact and remaining time.

Do not generate a generic question dump before building the target and story maps. Read `references/prep-pack.md` and `references/question-mapping.md`.

## Stage behavior

- Recruiter or first screen: positioning, timeline, motivation, communication, and basic fit.
- Hiring manager: ownership, judgment, collaboration, impact, and role-specific depth.
- Project or technical discussion: architecture, product logic, alternatives, evaluation, failure modes, and defensible implementation detail.
- Case or open problem: assumptions, structure, decisions, metrics, risk, and adaptation.
- Final or panel: unresolved concerns, consistency, maturity, values in action, and questions for the team.

These are starting hypotheses. The actual JD, company research, interviewer, and prior feedback take priority.

## Output rules

Make every prepared claim traceable to allowed `C###` items. Label predictions as inference. Match the requested interview language. End with one concrete next drill, usually the weakest high-priority question or riskiest flagship story.