ChatDOC Studio--KnowledgeMate

Create and operate ChatDOC Studio knowledge bases through pd_router using a Bearer API key and JavaScript helpers. Use when Codex needs to upload one or more PDF/DOC/DOCX files, skip failed files without aborting the whole job, create a knowledge base from successful uploads, or call the ChatDOC Studio knowledge-base skill endpoints through `https://platform.paodingai.com/openapi/chatdoc-studio/...`.

Run a JavaScript workflow that uploads local PDF/DOC/DOCX files to ChatDOC Studio through PDRouter, waits for successful parsing, and creates a knowledge base from the successful uploads in one step.

This is suitable for creating knowledge bases from local files or folders, then using the returned library_id for document listing, retrieval, grep, syllabus reading, and block reading.

Installation

npx skills add PaodingAI/skills

For a local Codex installation, place this skill directory under:

$CODEX_HOME/skills/chatdoc-studio-knowledgemate

Usage

Create a knowledge base from files:

node scripts/knowledge-mate.mjs upload-and-create \
  --name "Quarterly KB" \
  --file ./docs/a.pdf \
  --file ./docs/b.docx \
  --file ./docs/c.pdf

Create a knowledge base from a folder recursively:

node scripts/knowledge-mate.mjs upload-and-create \
  --name "Quarterly KB" \
  --dir ./docs

Query an existing knowledge base:

node scripts/knowledge-mate.mjs retrieval --library-id <id> --query <text>
node scripts/knowledge-mate.mjs list-documents --library-id <id>
node scripts/knowledge-mate.mjs stats --library-id <id> --upload-id <id>
node scripts/knowledge-mate.mjs grep --library-id <id> --upload-id <id> --pattern <text>
node scripts/knowledge-mate.mjs read --library-id <id> --upload-id <id> --offset <n>
node scripts/knowledge-mate.mjs read-syllabus --library-id <id> --upload-id <id>

Execution Constraints

• You must invoke scripts/knowledge-mate.mjs directly. Do not reimplement the PDRouter or ChatDOC Studio API flow yourself during normal use.
• Route every request through PDRouter at /openapi/chatdoc-studio/....
• Authenticate only with Authorization: Bearer <PAODINGAI_API_KEY>.
• Do not call chatdoc_studio directly.
• Do not add internal X-PD-* signature headers in this skill.
• Do not expose or use separate upload-only or create-only commands; use upload-and-create for new knowledge bases.
• Only inspect or modify the script implementation when the script itself is unavailable, failing, or needs a fix.
• The behavior contract below explains what the script does, what it outputs, and when to use it. It is not a manual checklist for the model to imitate step by step.

When to Use

• Use this skill when the user wants to create a ChatDOC Studio knowledge base from local PDF, DOC, or DOCX files.
• Use this skill when the user provides a folder and asks to build a knowledge base from the documents inside it.
• Use this skill when the user wants knowledge retrieval, document search, document stats, syllabus inspection, or reading specific document blocks from a ChatDOC Studio knowledge base.
• Use this skill when the workflow must skip failed or unsupported files without aborting the entire upload batch.
• Use this skill only for PDRouter-backed ChatDOC Studio skill APIs, not for direct ChatDOC Studio internal APIs.

Environment Variables

• PAODINGAI_API_KEY: Required. The Bearer API key for PDRouter. If it is missing, the script fails immediately and tells the user to create a Bearer API Key in pdrouter, then export it before retrying.

The PDRouter base URL is fixed in the script as https://platform.paodingai.com. The service code is fixed in the script as chatdoc-studio; users do not need to set base-url or service-code environment variables.

The API key can be obtained from the PDRouter platform.

Default Behavior and Optional Settings

• Upload concurrency is fixed at 5.
• Supported upload file types are .pdf, .doc, and .docx.
• --dir is scanned recursively.
• --file and --dir can be mixed in one command.
• Duplicate file paths are removed before upload.
• Unsupported file types are not uploaded and are reported as skipped failures.
• If more than 300 supported files are found, the script stops before uploading and asks the user to clean the folder to at most 300 PDF/DOC/DOCX files.
• If every file fails to upload or parse, knowledge-base creation is skipped and the script exits with a non-zero status.

Script Behavior

1. Read PAODINGAI_API_KEY from the environment and use the fixed PDRouter base URL https://platform.paodingai.com.
2. Expand all --file inputs and recursively scanned --dir inputs into one de-duplicated file list.
3. Count supported .pdf, .doc, and .docx files before uploading. If the count is greater than 300, fail immediately without uploading anything.
4. Upload supported files through POST /openapi/chatdoc-studio/knowledge-base/upload with concurrency 5.
5. Skip unsupported files, upload failures, parse failures, and per-file backend errors without aborting the whole batch.
6. Call POST /openapi/chatdoc-studio/knowledge-base only with successful upload_ids.
7. Print JSON output. Successful API responses are unwrapped to their data payload. Failed requests print a clear error reason and include the backend response when available.
8. After upload-and-create returns a library_id, use the read/query commands for retrieval, listing documents, stats, grep, syllabus reading, or block reading.

Resources

• Run scripts/knowledge-mate.mjs instead of rewriting the HTTP client by hand.