How does Session Wrap-Up know when to prompt me?

The skill monitors specific intent signals like 'wrapping up,' detects when you're nearing the 180K token limit, or identifies unchecked items in your active project plans.

What kind of data does it help capture?

It helps capture progress on active plans, draft architectural decisions (ADRs), and 'lesson-worthy' insights from recent commits that haven't been documented yet.

Can I use this with LLMs other than Claude?

While the skill is optimized for Claude Code, the underlying knowledge base it builds is platform-agnostic and designed to persist across any LLM assistant.

Will it prompt me twice if I use a Stop hook?

No, the skill uses a coordination flag mechanism to ensure that if a session summary has already been triggered by a platform hook, it avoids double-prompting.

Session Wrap-Up

Name: Session Wrap-Up
Author: technomensch

bytechnomensch

•

Productivity & Workflow

Captures project context, decision logs, and lessons learned to ensure seamless continuity between development sessions.

Session Wrap-Up is an intelligent continuity tool designed for Claude Code that preserves institutional memory by prompting users to summarize work at key milestones. It monitors context limits, active plans, and architectural decision records (ADRs) to identify uncaptured progress, preventing information loss between sessions. By coordinating with platform-specific hooks, it ensures that critical insights and pending tasks are documented without redundant prompting, allowing developers to pick up exactly where they left off with any LLM assistant.

Key Features

01Intelligent trigger detection based on user intent and context window limits

02Automated identification of uncaptured lesson-worthy code changes

03Active monitoring of draft Architecture Decision Records (ADRs) and project plans

044 GitHub stars

05Contextual awareness of existing snapshots to prevent duplicate prompts

06Cross-session persistence coordination using temporary status flags

Use Cases

01Ensuring project progress is summarized before reaching the 180K token context limit

02Managing transitions between different LLM sessions while maintaining a platform-agnostic knowledge base

03Capturing institutional memory and architectural decisions during complex development sprints

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add technomensch/knowledge-graph session-wrap

For use in Claude.ai and ChatGPT

Key Features

01Intelligent trigger detection based on user intent and context window limits

02Automated identification of uncaptured lesson-worthy code changes

03Active monitoring of draft Architecture Decision Records (ADRs) and project plans

044 GitHub stars

05Contextual awareness of existing snapshots to prevent duplicate prompts

06Cross-session persistence coordination using temporary status flags

Use Cases

01Ensuring project progress is summarized before reaching the 180K token context limit

02Managing transitions between different LLM sessions while maintaining a platform-agnostic knowledge base

03Capturing institutional memory and architectural decisions during complex development sprints

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add technomensch/knowledge-graph session-wrap

For use in Claude.ai and ChatGPT