- Home
- Skills
- Technical Documentation
- medium digest
medium digest
Identifies reputation-building blog ideas from your Claude Code history and drafts professional technical posts.
Free
Works with the AI Agents you already use
medium digest
Example session with this skill installed
check my work for blog ideas --all
- Read your context and instructions
- Compiled the medium digest
I have scanned 142 sessions across 8 projects from your entire history. While most sessions involved routine maintenance, I found 4 significant engineering narratives worth sharing.
Technical Blog Candidates
| Candidate Title | Reasoning |
|---|---|
| The Ghost in the Event Loop | A deep dive into a subtle race condition in the payment-gateway service that only manifested under specific CPU throttling. |
| Migration Without Downtime | Your approach to the zero-downtime schema refactor for the user-profiles table, focusing on the dual-write period. |
| Why We Deleted 2k Lines of "Clean" Code | The architectural decision to revert a complex abstraction in the auth-flow project in favor of simple, flat logic. |
| Diagnostic Tooling for Legacy Nets | The custom CLI tool you built to trace requests through the legacy monolith, which reduced MTTR by 40%. |
Next steps
- Select a candidate to generate a full Markdown draft and LinkedIn teaser.
- Provide any internal design docs related to these sessions to improve technical accuracy.
- Review the sanitized draft in
~/Documents/medium-digest/drafts/for sensitive data.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop letting your best engineering wins get lost in closed terminal sessions. This skill identifies the high-signal work buried in your local Claude Code transcripts and drafts professional technical content that builds your reputation as a distinguished engineer. It ignores the trivial fluff of daily coding to find the gnarly bug fixes, architectural tradeoffs, and production migrations that a senior technical audience actually wants to read.
What it does
- Smart Session Scanning walks your entire local history across all projects to identify high-impact technical work.
- Selective Curation filters out daily noise to surface 3-7 candidates with real architectural depth or counterintuitive lessons.
- Voice-Matched Drafting generates full Markdown posts using your specific narrative style and "Act 1-4" structure.
- Visual Asset Prep creates hero images and technical diagrams while adhering to platform-specific format constraints.
- Social Distribution prepares ready-to-post LinkedIn teasers that hook readers without sounding like an advertisement.
How it works
- Run the
scan_sessions.pyscript to index recent work without dumping full, bulky transcripts into context. - Review the shortlisted candidates and select the technical win that has the most teaching potential.
- Review the generated local Markdown draft and images, which are automatically sanitized of internal hostnames and secrets.
Frameworks & tools
Python for session history parsing, Markdown for content delivery, and local JSON for state management.
Why this beats prompting it yourself
Generic prompts produce "dev logs" that no one reads. This skill applies a specific "distinguished engineer" judgment layer, forcing the output to focus on transferable lessons and architectural reasoning rather than just documenting what you did.
Use cases
- Converting a difficult production incident into a "lessons learned" post.
- Turning a complex database migration into a technical guide for other staff engineers.
- Scaling your engineering brand by automating the "what should I write about" decision.
Known limitations
This skill only reads local Claude Code transcripts and cannot access external documentation unless a retrieval tool is provided. It never publishes directly to any platform.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Technical Documentation.
Requires local Claude Code session history in JSONL format. Does not automate publishing; user must manually post to Medium/LinkedIn. Sanitization is best-effort and requires a final human check.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 13 days ago
- Free with an Agensi account
- Software
Git