daily-takeaway

Distills today's Claude conversations and memory files into shareable social media insights.

1|Updated Nov 25, 2021
One-click install
npx skills add https://github.com/rstagi/dotfile --skill daily-takeaway-rstagi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: daily-takeaway
Source: https://github.com/rstagi/dotfile/tree/main/.claude/skills/daily-takeaway
Command: npx skills add https://github.com/rstagi/dotfile --skill daily-takeaway-rstagi

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? At the end of a working day it is hard to remember which moments were genuinely worth sharing publicly. This Skill scans your day's Claude conversation transcripts and memory updates, then surfaces 1–2 candidate insights you could post on X, LinkedIn, or a blog — and honestly tells you when nothing is worth posting. ## Core Features & Use Cases - Daily signal gathering: Scans memory files modified today and parses JSONL conversation transcripts (filtering out tool calls and system noise) to find notable moments. - Distillation rubric: Filters candidates for surprising, concrete, hook-worthy insights while discarding routine work and regrettable hot takes. - Multi-platform drafts: Produces X, LinkedIn, and blog-length variants of each insight, with source citations so you can verify framing before publishing. - Use Case: At 6pm you ask for an end-of-day takeaway; the Skill finds that a debugging session revealed a counterintuitive root cause, drafts three platform-ready posts about it, and appends them to ~/dotfile/insights/2024-06-01.md. ## Quick Start Ask the assistant to run the daily-takeaway skill and tell you what from today's work is worth posting on social media.

Frequently Asked Questions about daily-takeaway

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate social media posts from my Claude conversations?

Run the daily-takeaway skill at the end of your workday. It scans today's JSONL conversation transcripts and memory files, applies a quality rubric, and drafts X, LinkedIn, and blog-length variants of 1-2 insights with source citations.

How does the skill find today's conversation transcripts?

It reads JSONL files in the Claude projects directory, filtering entries by a UTC timestamp cutoff covering the last 30 hours to handle timezone boundaries. Only user and assistant text content is extracted; tool calls and system messages are skipped.

What happens if nothing worth posting happened today?

The skill writes an honest low-signal-day note to the output file with a one-sentence summary of what the day actually contained. It does not fabricate insights to fill the slot, since forced posting is considered worse than silence.

Does the skill require jq to parse transcripts?

jq is the preferred tool for filtering JSONL transcript lines by timestamp and message type. If jq is unavailable, the skill falls back to python3 with json.loads processed line by line.

Can I run the daily takeaway more than once per day?

Yes. If the output file for today already exists, the skill appends a new timestamped run section instead of overwriting. This supports re-running later in the evening after additional conversations.