feishu-deep-research_skill

Overview

This skill automates deep research and generates a structured Feishu document uploaded to a Feishu cloud folder. It runs a three‑phase workflow: user confirmation, autonomous multi‑round research with integrity checks, and direct Feishu REST API file import using exec + curl. The output includes a timestamped Markdown report, source list, and all upload/import artifacts (tenant token, file_token, ticket, job_status, doc_url).

How this skill works

After an initial confirmation step, the skill performs time‑layered searches (historical, recent, latest) at a controlled rate, enforces minimum source and time coverage by depth level, and runs automatic completeness checks. It assembles a Markdown research report with a mandatory data timeliness statement and source list, then uses curl calls (tenant_access_token → upload file → create import task → poll status) to import the file into Feishu cloud space. Every API step must be reported explicitly.

When to use it

• Investment research needing verifiable provenance and timeliness
• Industry or competitor analysis that requires multi‑source aggregation
• Technical due diligence or market scouting with strict audit trail needs
• When you must produce a Feishu document with documented API upload artifacts
• Any scenario that requires automatic completeness checks before report delivery

Best practices

• Always complete the startup confirmation to set date range and search rate
• Choose depth level (basic/standard/deep) to match required source count and time coverage
• Keep Markdown files under recommended size (≈100KB) or split by chapter
• Provide correct Feishu folder token and app credentials or the process will halt
• Do not bypass curl steps with wrapper libraries; the skill requires explicit exec + curl calls

Example use cases

• Generate a deep investment report on a listed company and upload the document to a team Feishu folder
• Produce quarterly competitor surveillance summaries with source URLs and timeliness declaration
• Run technical landscape research for a product team and produce a documented Feishu report for stakeholders
• Create an industry trend brief that includes at least three source types and automatic completeness validation