Daily Brief Bot: Automate Your Morning Reading Digest

Where this fits

Every morning you open Discord, hoping the research you asked for actually landed—but instead you find 40 unread articles and no summary. You want a bot that reads the latest posts from your favorite niche community, filters them against your personal knowledge vault, and sends you exactly three links worth your attention, in the tone you prefer, at a time you choose. That's the daily-brief-bot skill: it turns a vague morning-routine wish into a working, tunable automation.

Why agents benefit

• Filters by personal context: The bot compares new articles against a local vault (notes, memories, skills) to select only items you haven't already seen and that match your active projects—no generic keyword scraping.
• Owns a single prompt template: All output style (tone, length, count) lives in one prompt-template.md file, so an agent can tweak the whole behavior by editing one file instead of hunting through code.
• Validates its own output: A verify.mjs step cross-checks that every link in the summary actually came from the collected posts, preventing hallucinated or broken URLs.
• Handles the full lifecycle: Data collection (GraphQL), candidate ranking, Discord DM sending, and even capturing user replies—so agents can respond to follow-up instructions like "more technical" or "send at 7am."
• Self-describing triggers: Natural-language commands ("daily brief bot 2번 해보자" or "edit the tone") map directly to agent actions, making the skill easy to embed in an existing assistant.

Practical scenarios

Morning briefing on a niche community

You run a bot that pulls new posts from a Bettermode community (like gpters.org), filters them against your Obsidian vault, and DMs you a numbered list every day at 10:00. The skill includes the exact collection, filtering, and sending pieces.

Tuning the output without touching code

You want the brief to be shorter, more casual, and include only two items. Instead of editing Python scripts, you ask the agent to "make it more casual and shorten to 2 items." The agent edits prompt-template.md and the dispatch updates automatically.

Interactive follow-up on a single item

After receiving the brief, you reply "3번 자세히" to get a deeper summary of item 3. The capture script reads your reply, and the agent can re-run the pipeline or fetch more context from the vault—turning a one-shot digest into a conversation.

Add it to your agent workflow

Install the skill into your agent environment (requires Node.js and a remote SSH target):

npx skills add bam-bam-2/solo-skills --skill daily-brief-bot

Then point your agent to the project folder and start the daily run:

ssh remote-host "cd ~/Projects/daily-brief && ./run.sh"

The skill includes a launchd config for scheduling, so you can have it fire at 10:00 every day without manual intervention.

Compare before adopting

Before committing, compare against alternatives:

• Quality signals: Does the project have recent commits, issue responses, or a clear README that explains failure modes?
• Maintenance freshness: Check the GitHub repo for activity in the last 3 months—a stale skill that relies on a specific API version may break quickly.
• Alternatives: There are generic RSS-to-Discord bots, but they lack the vault-based filtering and the single-file prompt editing this skill provides.
• Workflow fit: The skill assumes you have a remote machine with SSH, a Discord bot token, and a personal vault (e.g., Notion, Obsidian). If those don't exist, you'll need to adapt.

Why it is worth tracking

This skill demonstrates a pattern that most agent automations miss: combining external data collection with a personal knowledge graph and allowing one small prompt file to control the entire user experience. It has 11 stars and a focused scope, making it a reliable reference for building your own digest bots. When you need a scheduled, context-aware notification pipeline, check the project's commit history and recent adjustments to see if it matches your community and storage setup. For a quick overview, see the skill's page at OpenAgentSkill.