WeCom Daily Report Assistant

Problem

WeCom daily reports often arrive as free-form text and are hard to track by project, owner, and date. This skill parses report text into structured rows and writes them to a WeCom Smartsheet, reducing manual cleanup and duplicate entry.

How It Works

The skill works against the target Smartsheet fields:
- Read table structure: calls smartsheet_get_fields to get single-choice options for Project Name, such as Research, Testing, and Development, instead of hard-coding project names.
- Parse content: extracts employee name, project, today's work, issues, and time spent; it prefers the date in the report and falls back to the current date.
- Map identities: if the report lacks a clear name, config.json can map WeCom user IDs to real names; if no mapping exists, the user ID is used directly.
- Deduplicate writes: multiple projects become separate records; for the same person on the same day, local record_ids.json is used to delete the previous record, insert the new record, and update the stored record ID.

Limits

The target Smartsheet must include fields such as project name, employee name, date, today's work, issues, and time spent. Because existing remote records cannot be queried directly, deduplication depends on the local record-ID file and should not be deleted casually. The issue field is currently written only to the first record, so cross-project issues may need manual cleanup.