| name | offerpilot-cursor-adapter |
| description | Use the OfferPilot skill pack for resume optimization, China-first JD fit diagnosis, targeted rewrites, structured evaluation, mock interview prep, product research, application tracking, outreach, and cover letters in Cursor or other repository-local coding agents. Trigger on short intents like "按照 offerpilot 优化简历", "用 offerpilot 做 JD 匹配", or "/offerpilot ...". |
Cursor Adapter
This adapter is a thin wrapper over the platform-agnostic OfferPilot skill pack.
Read Order
../../README.md
../../WORKFLOW.md
../../INPUTS.md
../../OUTPUTS.md
../../PROMPTS.md
- task-specific document:
- China-first JD matching:
../../JD_MATCHING.md
- resume diagnosis:
../../RESUME_DIAGNOSIS.md
- profile datastore assembly:
../../DATASTORE.md
- structured or batch evaluation:
../../EVALUATION.md
- mock interview:
../../MOCK_INTERVIEW.md
- product research:
../../PRODUCT_RESEARCH.md
- application tracking:
../../TRACKER.md
- LinkedIn outreach:
../../OUTREACH.md
../../scripts/README.md when local helper scripts are useful
When To Use
- resume diagnosis
- resume optimization
- China-first JD fit diagnosis
- targeted resume rewriting
- structured and batch job evaluation
- job discovery and recommendation
- mock interview question generation and evaluation
- product research for interview preparation
- application tracking and follow-up review
- LinkedIn outreach message generation
- cover letter generation
- multilingual application material cleanup
Short Triggers
Treat these as direct triggers for this skill:
按照 offerpilot 优化简历
用 offerpilot 做 JD 匹配
/offerpilot 优化简历
/offerpilot 根据我简历推荐岗位
/offerpilot 结构化评估
/offerpilot 模拟面试
/offerpilot 产品研究
/offerpilot 外联消息
/offerpilot 简历诊断
/offerpilot 求职信
/offerpilot 申请追踪
/offerpilot 批量评估
When these short triggers appear, do not ask for a long setup prompt first. Start from the read order, then request only missing required inputs.
Cursor-Specific Notes
- prefer local repository files as inputs
- use the rules in
../../README.md, ../../INPUTS.md, and ../../OUTPUTS.md as the source of truth instead of repeating platform-local variants
- run
python3 skill-pack/scripts/extract_text.py "path/to/file" when a .pdf or .docx source cannot be read reliably
- use validation scripts when they reduce risk:
validate_inputs.py, validate_outputs.py, validate_profile_store.py, and validate_aliases.py