Game Center Tracking CSV

Problem

Game Center tracking docs often drift because inputs are scattered: oper_id, game_id, and activity_id lack a source, page_name and element_level naming is inconsistent, paired exposure/click events are missing, and CSV files break Chinese text in Excel. This skill targets tracking doc generation, completion, and review, turning Figma, PRDs, screenshots, and existing CSVs into a deliverable table without inventing business facts.

How It Works

• Reads references/tracking_spec.md, references/example_patterns.md, and assets/tracking_template.csv to confirm column order, required fields, and extension-field semantics.
• Extracts pages, modules, elements, interactions, and events from the source material, then builds an event matrix for page show/hide, module exposure, element exposure/click, playback, scroll, and download.
• Populates page_name, module_name, element_name, element_level, game_id, activity_id, adtag, and ext1-ext25 according to the spec.
• Leaves unknown fields empty or marks them as “to confirm” rather than fabricating oper_id or business parameters.
• Uses scripts/tracking_csv_tool.py to generate a template, write output.csv, and run validate, then revises errors and warnings.

Boundaries

The skill relies on the stated spec and reference patterns, so it works best when Figma, PRD, screenshots, or an old CSV are available. Missing real oper_id, entry source, dynamic parameters, or navigation targets should remain pending. Complex permission states, failure states, and cross-page state machines still need human source checks.