Hacker News Weekly Digest Automation

agent_builder_prompt.md

Fetch the top 10 latest news from Hacker News and create a professional HTML newsletter.

The newsletter should have:

• A visually appealing design with Hacker News orange theme (#ff6600)
• Full HTML structure with inline CSS styling
• Each story formatted with:
  • Story title as H3 header
  • Timestamp showing when it was posted
  • Brief description/snippet from the article
  • Clickable link to the HN discussion thread
• Professional styling with:
  • Clean typography (Arial/sans-serif)
  • Card-based layout for each story
  • Proper spacing and padding
  • Border accents using HN orange color
  • Mobile-friendly responsive design
• Header section with newsletter title and date
• Footer with branding

Send the HTML newsletter to my email address.

clean_json_response.py

# Extract LLM output
text = items[0]["json"]["content"]["parts"][0]["text"]

# Remove escaped newlines
text = text.replace("\\n", "")

# Remove literal newlines
text = text.replace("\n", "")

# Clean whitespace
text = text.strip()

# Return ONLY one item but attach it to the first input
return [{
    "json": {
        "html": text
    },
    "pairedItem": {
        "item": 0
    }
}]

clean_schema.py

# extract specific fields
# store in dict & pair it with original - n8n specific
import json

output = []

for index, item in enumerate(items):
    d = item.get("json", {})

    cleaned = {
        "title": d.get("title", ""),
        "url": d.get("url", ""),
        "author": d.get("author", ""),
        "points": d.get("points", 0),
        "comments": d.get("num_comments", 0),
        "created_at": d.get("created_at", "")
    }

    output.append({
        "json": cleaned,
        "pairedItem": {
            "item": index
        }
    })

return output

parse_schema_to_json.py

# parse the string into actua json dict
import json

stories = []

for item in items:
    d = dict(item.get("json", {}))

    stories.append({
        "title": d.get("title", ""),
        "url": d.get("url", ""),
        "author": d.get("author", ""),
        "points": d.get("points", 0),
        "comments": d.get("comments", 0),
        "created_at": d.get("created_at", "")
    })

# Pretty-print for humans
pretty = json.dumps(stories, indent=2)

return [{
    "json": {
        "stories_raw": stories,  # for eyes
        "stories": pretty        # for the LLM
    }
}]

prompt.md

You will receive a JSON array called "stories".

Your job: Generate a clean, email-ready HTML digest.

Layout:

Hacker News Digest

Short intro sentence...

Then for each story:

TITLE

Author: AUTHOR

Points: POINTS  |  Comments: COMMENTS

Published: CREATED_AT

Read more

Rules:

• Output ONLY HTML.
• NO Markdown.
• NO "\n".
• NO "\".
• Use actual HTML line breaks or paragraphs, not text newlines.
• Every new line must be a real newline character, not "\n".
• The output must be valid HTML.
• After each story add a seperator using -----------

Published Field formatting rule:

When displaying the Published date/time, use this transformation:

• Extract the date as the part before "T". Example: "2020-10-23".
• Extract the time as the HH:MM portion only. Remove seconds and remove the trailing "Z". Example: "19:26:35Z" becomes "19:26".

Display it in this format:

Published: DATE at TIME

Here is the JSON array:

{{ $json.stories }}