langflow-document-summary/
├── SKILL.md
├── scripts/
│   └── run_flow.py
└── references/
    └── api-contract.md

---
name: langflow-document-summary
description: Use when the user wants Codex to summarize a local document by calling the configured LangFlow document-summary flow and saving a validated Markdown or JSON result.
---

# LangFlow Document Summary

Use this skill when the user asks to summarize, extract, or structure a document through the team's LangFlow document-summary flow.

## Workflow

1. Locate the document the user wants processed.
2. Extract plain text if the input is PDF, DOCX, HTML, or Markdown.
3. Run `scripts/run_flow.py` with the extracted text.
4. Validate that the response includes `summary`, `key_points`, and `open_questions`.
5. Save the result under the requested output folder.
6. Report the saved path and any validation concerns.

## Environment

Required variables:

- `LANGFLOW_BASE_URL`
- `LANGFLOW_API_KEY`
- `LANGFLOW_FLOW_ID`

Read `references/api-contract.md` if the response shape is unclear.

# API Contract

The flow receives one text input and returns a structured response.

Required input:

- `input_value`: full document text
- `input_type`: `chat`
- `output_type`: `chat`

Expected output:

- `summary`: short prose summary
- `key_points`: list of important claims or facts
- `open_questions`: list of unclear or missing items

import json
import os
import sys
from pathlib import Path

import requests


def require_env(name: str) -> str:
    value = os.environ.get(name)
    if not value:
        raise RuntimeError(f"Missing required environment variable: {name}")
    return value


def run_flow(input_text: str) -> dict:
    base_url = require_env("LANGFLOW_BASE_URL").rstrip("/")
    flow_id = require_env("LANGFLOW_FLOW_ID")
    api_key = os.environ.get("LANGFLOW_API_KEY")

    headers = {"Content-Type": "application/json"}
    if api_key:
        headers["x-api-key"] = api_key

    payload = {
        "input_value": input_text,
        "input_type": "chat",
        "output_type": "chat",
    }

    response = requests.post(
        f"{base_url}/api/v1/run/{flow_id}",
        headers=headers,
        json=payload,
        timeout=120,
    )
    response.raise_for_status()
    return response.json()


def main() -> None:
    if len(sys.argv) != 3:
        raise SystemExit("Usage: run_flow.py INPUT_TEXT_FILE OUTPUT_JSON_FILE")

    input_path = Path(sys.argv[1])
    output_path = Path(sys.argv[2])

    result = run_flow(input_path.read_text(encoding="utf-8"))
    output_path.write_text(json.dumps(result, indent=2), encoding="utf-8")
    print(output_path)


if __name__ == "__main__":
    main()

Use the LangFlow document summary skill on this PDF and save the result.

If it transforms data inside an AI pipeline, put it in LangFlow.
If it tells Codex how to operate, validate, and file the work, put it in a skill.
If it must be exact every time, put it in a script and let the skill call it.