Feishu Report Summary

Use this skill when the user wants a summary of Feishu 汇报 data and the source link is under oa.feishu.cn/report/.... The bundled script calls the Feishu Report API directly, using the same credentials already stored in ~/.openclaw/openclaw.json.

Workflow

1. Run scripts/fetch_report_tasks.js for the target day or range.
2. Prefer --format json when you will analyze the results in-model.
3. Prefer --format markdown when the output will be reviewed or copied directly.
4. Summarize with the structure in references/summary-template.md unless the user asks for a different format.
5. If the user wants the digest published back to Feishu, use existing feishu_doc or feishu_chat tools after the summary is drafted.

Quick Start

Fetch today's local-calendar reports as JSON:

node scripts/fetch_report_tasks.js --format json

Fetch one day and cap the export:

node scripts/fetch_report_tasks.js --date 2026-03-14 --max-items 50 --format json

Fetch one rule for the last 7 days and render Markdown:

node scripts/fetch_report_tasks.js --days 7 --rule-name "研发团队工作日报" --format markdown

Write the export to a file for later review:

node scripts/fetch_report_tasks.js --date 2026-03-14 --output /tmp/feishu-report.md

Script Notes

• The script reads Feishu credentials from ~/.openclaw/openclaw.json by default.
• Use --account-id when multiple Feishu accounts are configured.
• --date, --start-date, --end-date, and --days use the machine's local timezone.
• Use FEISHU_APP_ID and FEISHU_APP_SECRET only when you need to override config values.
• The script resolves --rule-name to a rule ID before querying tasks so the final query stays precise.

Expected Output

The JSON payload contains:

• query range metadata
• resolved filters
• task counts grouped by rule and reporter
• normalized task entries with field names and pretty-printed field values

Use that payload to produce a concise digest that highlights:

1. common themes
2. completed work
3. blockers or risks
4. next actions
5. people or teams that need follow-up