Does it support different output formats?

Yes, you can choose between concise bullet points, detailed markdown reports with evidence, or high-level executive summaries, and optionally save them to a file.

Do I need to convert my .cast files first?

Yes, this skill is optimized for .txt files. You should use the `/asciinema-tools:convert` command to transform your raw .cast recordings into text before summarizing.

What kind of focus topics can I provide?

You can provide any specific topic such as 'errors,' 'ML training results,' 'database migrations,' or 'performance metrics' to guide the AI's search patterns.

Can it handle files larger than Claude's context window?

Yes, the skill uses strategic sampling (head, middle, tail) and grep-based keyword extraction to analyze files as large as several gigabytes without exceeding memory limits.

Asciinema Session Summarizer

Name: Asciinema Session Summarizer
Author: terrylica

byterrylica

•

生産性とワークフロー

Analyzes large terminal recordings using iterative sampling and AI-powered extraction to provide deep-dive insights and reports.

This skill enables Claude to perform intelligent, iterative analysis of massive terminal recordings converted from .cast to .txt format. By utilizing strategic sampling—targeting the beginning, middle, and end alongside keyword-driven extraction—it overcomes context window limitations for files exceeding 1GB. Users can specify focus areas such as error debugging, ML training results, or architectural decisions, and Claude will use interactive questioning to drill down into specific findings, ultimately generating structured markdown reports or executive summaries of long development sessions.

主な機能

01Interactive 'Drill Down' mode using multi-select questions for guided session exploration.

02Configurable analysis depth levels (quick, medium, deep) to balance speed and detail.

03Domain-specific keyword libraries for ML research, software development, and DevOps workflows.

04Handles massive recordings (1GB+) through strategic file sampling and keyword-targeted extraction.

0524 GitHub stars

06Automated synthesis of findings into structured Markdown reports with specific line-number evidence.

ユースケース

01Analyzing long-running ML training sessions to identify convergence issues or specific breakthroughs.

02Post-mortem debugging of complex DevOps automation tasks and terminal-based workflows.

03Summarizing multi-day collaborative coding sessions to extract key decisions and next steps.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add terrylica/cc-skills summarize

For use in Claude.ai and ChatGPT