What is mid-conversation compaction?

Compaction occurs when the context window is full, forcing the AI to summarize or drop earlier parts of the conversation. This skill helps prevent it.

Why is the TaskOutput tool discouraged?

TaskOutput reads full agent transcripts into the main context window, which can exceed 70,000 tokens, leading to context overflow and reduced reasoning quality.

Will I still see the results of my background tasks?

Yes, you will receive a concise summary of the task execution instead of the raw, verbose logs that consume excessive tokens.

How does this skill change my workflow?

It encourages using the Task tool in synchronous mode (run_in_background=false), which returns a summary rather than a full transcript dump.

Context-Efficient Task Management

Name: Context-Efficient Task Management
Author: scooter-lacroix

byscooter-lacroix

Productividad y Flujo de Trabajo

Optimizes agent performance by preventing context window flooding and token overflow during complex task execution.

Acerca de

The no-task-output skill is a performance-focused utility designed to maintain the integrity of the Claude Code context window. It enforces a strict rule against using the TaskOutput tool, which can dump over 70,000 tokens of agent transcripts into the main session, causing premature context compaction. By prioritizing synchronous task execution, this skill ensures that only concise summaries are returned to the main context, preserving token space for reasoning and long-term project history while maintaining agent isolation.

Características Principales

0 GitHub stars
Maintains agent context isolation for better reasoning
Prevents context window flooding from agent transcripts
Reduces token consumption in high-scale sessions
Enforces synchronous task execution patterns
Prevents mid-conversation context compaction

Casos de Uso

Managing complex multi-step agentic workflows
Improving performance of background task monitoring
Optimizing token usage during long development sessions

Acerca de

Características Principales

0 GitHub stars
Maintains agent context isolation for better reasoning
Prevents context window flooding from agent transcripts
Reduces token consumption in high-scale sessions
Enforces synchronous task execution patterns
Prevents mid-conversation context compaction

Casos de Uso

Managing complex multi-step agentic workflows
Improving performance of background task monitoring
Optimizing token usage during long development sessions