Evidence-first tools and a six-week learning system for Chinese campus recruiting.

中文完整指南

---

What is qiuzhao?

qiuzhao is an open-source system for Chinese graduate and internship applications. It connects role discovery, evidence-based fit ranking, resume tailoring and audit, application tracking, interview preparation, honest open-source contribution, and structured study.

The repository has two primary entrances:

Entrance	Use it when	Main outputs
Job-search Skills	You have job descriptions, a resume, or an application form	Ranked roles, tailored resume content, field mappings, interview packs
Six-week interview sprint	You are already applying but need stronger coding and fundamentals	42 daily plans, LeetCode reviews, CS fundamentals, Agent/RAG study, mock interviews

Important

qiuzhao does not invent experience or guarantee an offer. Claims must remain grounded in verifiable evidence. Login, CAPTCHA, declarations, and final submission always require the user's confirmation.

Included Skills

Skill	Purpose	Key output
job-fit-ranker	Apply hard eligibility gates and rank roles against resume evidence	Eligibility, fit score, evidence, gaps, official links
jd-resume-tailor	Tailor an existing resume to one target role without fabricating claims	Rewrites, keyword coverage, fact-check list
resume-auditor	Audit ownership, metrics, timelines, attribution, and evidence	Claim ledger, risky wording, corrected copy
application-form-helper	Map ATS fields and preserve the application event history	Field plan, conflicts, status, next action
interview-prep	Build JD-specific technical and behavioral preparation	Project deep dives, answer outlines, follow-ups, gap-repair plan
oss-contributor	Prepare a compliant contribution and preserve its actual state	Validated diff, confirmation gates, contribution record

Six-week interview sprint

The Six-week interview sprint is a GitHub-readable Obsidian Vault for candidates targeting AI Agent and RAG application-development roles. It assumes six hours of study per day and advances six tracks in parallel:

• Python programming
• LeetCode and live coding
• operating systems, networking, and databases
• Agent/RAG concepts and experiments
• project communication and mock interviews
• applications and daily review

Start with the Dashboard, Day 01, or the 42-day task cards.

Install

Install every Skill:

npx skills add luosir1123-ai/qiuzhao -g

Install one Skill:

npx skills add luosir1123-ai/qiuzhao --skill job-fit-ranker -g

Install from a local clone:

git clone https://github.com/luosir1123-ai/qiuzhao.git
cd qiuzhao
npx skills add . -g

To use the learning system, open 6周冲刺八股和力扣/ as an Obsidian Vault. Community plugins are optional; the core content uses standard Markdown and relative links.

Workflow

Official roles / JDs
        ↓
Eligibility and evidence-based fit ranking
        ↓
Resume tailoring → application-field consistency
        ↓
Interview preparation → gap repair → review
        ↖ honest OSS evidence and application history

The workflow keeps four categories separate:

• Evidence: facts supported by the resume, project artifacts, experiments, or official job descriptions.
• Inference: conclusions derived from evidence and labeled as such.
• Gaps: requirements with weak or missing evidence.
• Unknowns: information that was unavailable or still needs user confirmation.

Principles

• Apply cohort, degree, major, language, and location hard gates before scoring.
• Never turn internships, projects, research, or campus activities into a different experience category.
• This repository does not require committing or uploading a real resume and provides no hosted backend. Data handling still depends on the runtime and tools you choose.
• Never bypass CAPTCHA, identity checks, declarations, or final-submit confirmation.
• Treat coding submissions as evidence only after closed-book reproduction and explanation.

Development

python3 -m unittest discover -v
python3 scripts/validate.py
python3 scripts/check_private_data.py .
python3 "6周冲刺八股和力扣/scripts/validate_vault.py"

Examples must use fictional identities and example.com addresses. See AGENTS.md and CONTRIBUTING.md.

License

MIT. See THIRD_PARTY_NOTICES.md for upstream inspirations and license notices.