Gaokao Application Advisor Pro

Problem

The hard part of gaokao application planning is not listing schools, but mixing scores, ranks, years, provinces, and subject modes. A 580 score means different things across provinces, years, and batches, and directly outputting probabilities or “sure admission” can mislead users. This skill treats the workflow as a verifiable data pipeline: confirm parameters first, query reliable sources, and only then make reach, match, and safe recommendations.

How It Works

It separates sessions into data lookup, application planning, and specific questions. For data lookup, it prefers built-in scripts such as score-to-rank-lookup and regional-passing-scores, using province, year, subject class, score, and batch as inputs. Built-in alias mapping and exam-mode rules reduce errors from inputs like “Yue”, “Shenzhen”, or “physics track”. Year mismatches must be labeled explicitly, and future-year predictions or unverified probabilities are disallowed.

In the full planning flow, it moves through five stages: collect province, subject track, score, rank, restrictions, and adjustment preference; validate interest and profile via T1/T2 tables; clarify city, school tier, and graduate or employment goals through dialogue; then produce a draft table and final report. If Tencent Docs authorization is unavailable, it can fall back to a local Markdown checklist.

Limits

This skill is useful after scores are released or reasonably estimated, not for predicting unannounced batch lines, postgraduate admission rates, or employment outcomes. Postgraduate, employment, and prospect modules currently rely on simulated databases and must include source and warning labels. Outputs should avoid “sure admission” and other deterministic language, and all recommendations should be anchored to province, year, subject track, rank, and official sources.