SuperAPI AI Directory

gridcraft-clean-table

Data & analytics

View repository

Updated 11-04 23:02

About this skill

Table cleaning

Before executing the specific method, read Craft scenario commands and parameters, and execute according to the object selection, session, input replacement, and result checking within.

Task starting point

First read the materials, confirm the delivery purpose, target size or duration, modification scope, and required format. Details that can be determined from the task and materials should directly adopt reasonable defaults; only ask about missing conditions that would change the result.

When a visible editor is needed, hand off to the gridcraft-editor skill. Standalone installation:npx skills add full-aigc-skills/gridcraft-skills --skill gridcraft-editor. The currently added skill is a local candidate; before release, use the local companion plugin, and do not treat this installation command as proof of release.

Execution steps

  1. Keep the original table, and identify field types and primary keys in a working copy.
  2. Unify dates, units, and whitespace characters, and avoid converting IDs into numbers.
  3. Mark duplicates and missing values according to explicit rules, and do not fill them in by guessing.
  4. Summarize the row counts and anomalies before and after cleaning.

First confirm that the output is blank, then set the target primary key column as text before splitting columns, preserving leading zeros; by default, delete only completely identical whole rows, and leave primary key conflicts for review.

Tool implementation

Use open_workbook/new_workbook to manage workbooks; read_range to inspect, set_cell to edit, execute_command with complete commands; save_workbook to save.

First read the actual tool schema, then fill in the parameters; use the returned values for object or document IDs. Queried commands must be available on the current backend. Before saving, choose a new output path and keep the input. If the required algorithm, format, or command does not exist, explain the gap, and do not pass off a similar name as the real thing.

Result checking

  • Long IDs, leading zeros, and dates are not distorted.
  • Deleted or merged records can be mapped back to the original rows.
  • The formula area and data area are separated, and the output can be reopened.

Check the actual content against the user's goal, not just the process exit code. For local edits, inspect local details; for layout tasks, inspect the whole page; for time-based media, inspect the start, end, and key segments. For abnormal results, first check the current project to avoid repeatedly performing irreversible operations.

Delivery

Return the editable project, the actual derived files, and a brief explanation: what modifications were made, which steps were not executed due to current capabilities, and what requires the user's judgment. List only files that actually exist and have been checked. If the user subsequently reworks it, read this source project and make local modifications, preserving the content that has already been accepted.