A specialized skill for batch ingesting Feishu folders and attachments into reviewable, report-first knowledge artifacts for AI agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install feishu-knowledge-ingest
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install feishu-knowledge-ingest using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Feishu Knowledge Ingest?
Feishu Knowledge Ingest is a robust utility designed for Openclaw Skills that transforms messy Feishu directories or shared attachments into structured, high-quality data. It excels at classifying native Feishu documents versus uploaded files, ensuring that AI agents only learn from verified text extractions rather than making assumptions based on filenames.
By prioritizing a report-first architecture, this skill generates candidates for memory and detailed logs of successes and failures. This approach provides a controlled workflow for policy manual ingestion, directory learning, and general knowledge training without the risk of polluting primary memory files with unverified or poorly parsed data.
Feishu Knowledge Ingest Use Cases
- Ingesting large Feishu directories for AI model training or RAG (Retrieval-Augmented Generation) workflows.
- Converting corporate policy manuals and documentation stored in Feishu into structured JSONL formats.
- Auditing Feishu folders to identify permission-blocked or unparseable file types.
- Creating candidate memory files for human review before finalizing AI knowledge bases.
How Feishu Knowledge Ingest Works
- Input Resolution: The skill resolves the provided Feishu folder link or single attachment token to enumerate all accessible files.
- Manifest Creation: It generates a unique batch ID and builds a manifest tracking file names, types, and intended routing decisions.
- File Routing: Files are categorized into direct-read (native docs), download-and-parse (DOCX/PDF), or manual-review (images/PPTX).
- Text Extraction: The skill utilizes specialized parsers for .docx and .pdf files to extract raw text content.
- Artifact Generation: Successfully extracted data is written to kb-items.jsonl, while errors are logged to failed-items.jsonl.
- Batch Summarization: The process concludes by generating an ingest-report.md and a MEMORY.candidate.md for final review.
Feishu Knowledge Ingest Setup
To integrate this into your environment, ensure you have the necessary parser dependencies and run the local batch runner.
# Install required dependencies
pip install python-docx pypdf
# Execute a local test run
python run.py --input <FEISHU_URL_OR_TOKEN>
Feishu Knowledge Ingest Data Schema & Taxonomy
The skill produces four primary output artifacts to maintain a clean data taxonomy:
| File | Purpose | Key Fields |
|---|---|---|
kb-items.jsonl |
Validated knowledge | source_token, content_type, summary, confidence |
failed-items.jsonl |
Error tracking | failure_reason, error_detail, suggested_action |
MEMORY.candidate.md |
Memory staging | batch_header, grouped_summaries, source_refs |
ingest-report.md |
Executive summary | file_counts, routing_stats, recommended_actions |
Feishu Knowledge Ingest Advanced Features
- Strict Parsing Validation: Openclaw Skills native logic prevents the invention of text; if a parse fails, it is recorded with a concrete error reason.
- Granular Routing Rules: Automatically distinguishes between Feishu native formats (Wiki, Bitable) and uploaded binary attachments.
- Safety-First Memory: Uses a candidate system to prevent sensitive or unverified data from being automatically written to the primary MEMORY.md file.
- Multi-Parser Support: Includes modular helper scripts for DOCX and PDF text extraction that can be extended for other file types.
SKILL.md
Loading
Related Openclaw Skills
ACE Provisioning ManagementPartner API
A specialized tool for AI agents to manage Microsoft ManagementPartner provisioning and partner lifecycle operations via the Azure Resource Manager.
Skill Advisor
A comprehensive pre-installation assessment tool designed to evaluate the safety, maintenance, and technical requirements of Openclaw Skills.
Coldwallet Reference Tool
A comprehensive CLI-based reference utility for managing cold wallet protocols and navigating blockchain security standards.
Starbucks Skill
An AI-powered skill for retrieving Starbucks product menus, promotional offers, and store locations.
Baidu Content Censor
A powerful AI moderation tool for real-time text and image safety screening integrated via Openclaw Skills.
Hong Kong Bus & MTR ETA Agent
A high-performance transit tracking skill for real-time Hong Kong bus and MTR heavy rail arrival information via natural language queries.
Featured*
