# Career Os Router

> Use when a career request may require choosing, sequencing, or resuming one or more Career OS skills across profile, project, story, resume, target research, interview preparation, practice, live support, or debrief.

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

---


# Career OS Router

**REQUIRED BACKGROUND:** Load `career-state-protocol`, then load only the professional skill needed for the current step.

## Routing precedence

Apply this order to the latest user intent:

1. `Live` interview request.
2. `explicit mode` or named professional skill.
3. `scheduled interview` preparation, mock, or debrief.
4. `JD-specific` research, resume, or preparation.
5. `project or story` asset building.
6. `career profile` foundation or update.

Newer intent overrides older plans. Do not let a large prior context delay a live answer.

## Scope

Career OS covers candidate evidence, profile, project depth, story bank, resume content, target analysis, interview preparation, answer coaching, mock interviews, live assistance, and debrief.

It excludes job search, scraping, job ranking, application submission, outreach, cover letters, portfolios, LinkedIn, offer evaluation, compensation, coding test, algorithm, and written-exam workflows. Explain the boundary briefly and stop or use a separately available skill.

## Route table

- `career-profile-builder`: durable profile, import, conflict, capability, positioning.
- `project-deep-dive`: project evidence, Deep-Dive Map, architecture, product, evaluation, result, or ownership depth.
- `story-bank-builder`: story portfolio, coverage, answer-length variants, story strength.
- `jd-company-research`: supplied JD, company, product, role model, current research.
- `resume-builder`: master or targeted resume content and audit.
- `interview-prep`: stage-specific preparation pack, question map, story map, training plan.
- `answer-coach`: one answer, critique, rewrite, ladder drill, or coached retry.
- `mock-interview`: practice mode or uninterrupted formal simulation.
- `live-interview`: immediate in-interview answer.
- `interview-debrief`: fresh memory capture, transcript review, feedback, and write-back.

Read `references/routing-table.md` for compound flows and `references/intake-gates.md` for input requirements.

## Auto-chain policy

Use the minimum necessary skills and auto-chain until one of these confirmation gates:

1. `Deep-Dive Map scope` selection.
2. `high-risk Claim` confirmation or conflict resolution.
3. `formal artifact` publication: profile positioning, project dossier, Story Bank, resume, or final prep pack.

After confirmation, continue the chain without requiring the user to name the next skill. Do not preload or narrate all skills.

## Routing behavior

Inspect existing Career State before asking questions. Ask the highest-leverage missing input one question at a time unless fast mode is requested. For partial asset work, produce a provisional structure and gaps. For target-specific final output, enforce target gates. For live mode, answer first.

Default to content enrichment. Respect an explicit source-only request for that task. Preserve state and route newly discovered facts through Claim promotion.

