joes-toolkit

Analyze a single chunk of a long conversation, producing a deeply granular 8-section analysis with exact quotes, behavioral observations, and emotional arc tracking. Used by the conversation-summary skill to process chunks in parallel.

Synthesize multiple chunk analyses into a cohesive literary narrative that tells the complete story of a long conversation. Used by the conversation-summary skill after all chunk analyses are complete.

Transform long conversations into granular phase analyses and narrative synthesis. Use when a user provides conversation files (ChatGPT exports, interview transcripts, message logs, Slack exports, meeting notes) and asks to summarize, analyze, or extract insights. Triggers include requests to summarize conversations, analyze chats, process ChatGPT exports, summarize a directory of conversation files, or extract insights from long transcripts. Handles conversations too large for a single context window by chunking, parallel analysis, and narrative synthesis.

Guide for creating Claude Code plugins. This skill should be used when the user wants to "create a plugin", "write a plugin", "build a plugin", "scaffold a plugin", "set up plugin structure", "configure plugin.json", "add commands to a plugin", "add agents to a plugin", "add hooks to a plugin", "add MCP server to a plugin", "add LSP to a plugin", "organize plugin components", or needs guidance on Claude Code plugin architecture, manifest configuration, or plugin distribution.

Generate dark, dry tech humor using specific joke frameworks. Use when the user asks for a joke, wants something funny, asks to be entertained, requests memes, or says things like "make me laugh", "tell me a joke", "give me something dark", or "roast devops". Also use when humor would fit naturally into a conversation about distributed systems, monitoring, infrastructure, automation, AI, Kubernetes, or bureaucratic software.