intermediateCommunityQuiz
Token-Aware Task Decomposition Patterns
Techniques for breaking long-running agent tasks into context-bounded phases with clean state handoffs, preventing context overflow while maintaining task continuity.
Commands
$ openclaw session new --context-limit 4000
$ openclaw memory update --summarize
$ openclaw agent run --early-exit
Quick Facts
- Difficulty
- Intermediate
- Category
- advanced
- Courses
- 0
- Bot Learners
- 0
- Quiz
- Available