no-task-output
Never Use TaskOutput
When & Why to Use This Skill
This Claude skill optimizes agent performance and token efficiency by preventing context window flooding. It enforces a best-practice rule to avoid the 'TaskOutput' tool, which can dump over 70k tokens of raw transcripts into the main conversation. By switching to synchronous task execution, it ensures agent context isolation and prevents mid-conversation compaction, maintaining the quality and coherence of long-running AI sessions.
Use Cases
- Scenario 1: Managing complex multi-agent workflows where sub-tasks generate extensive logs that would otherwise overwhelm the main context window.
- Scenario 2: Optimizing token usage in long-form research or coding projects to prevent premature context reaching its limit.
- Scenario 3: Debugging and refining agent orchestration to ensure the main model receives concise summaries rather than raw, redundant execution data.
| name | no-task-output |
|---|---|
| description | Never Use TaskOutput |
| user-invocable | false |
Never Use TaskOutput
TaskOutput floods the main context window with agent transcripts (70k+ tokens).
Rule
NEVER use TaskOutput tool. Use Task tool with synchronous mode instead.
Why
- TaskOutput reads full agent transcript into context
- This causes mid-conversation compaction
- Defeats the purpose of agent context isolation
Pattern
# WRONG - floods context
Task(run_in_background=true)
TaskOutput(task_id="...") // 70k tokens dumped
# RIGHT - isolated context, returns summary
Task(run_in_background=false) // Agent runs, returns summary
Source
- Session where TaskOutput caused context overflow