self-challenge - Run dual-agent challenges to expand capabilities
Designs and executes capability challenges with Curriculum and Executor agents, then reviews outcomes and updates knowledge.
Tags
Updated: 2026-09-28self-evolutioncapability assessmentchallenge executiondual-agentknowledge managementskill development
Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Design capability challenges
- Execute challenge tasks
- Analyze challenge outcomes
- Assess capability changes
- Extract actionable learnings
- Update knowledge files
- Schedule follow-up challenges
Inputs
- User goals
- Self-reflection reports
- Existing skills
- Recommended skills
- Challenge specifications
- Reference resources
- Execution logs
Outputs
- Challenge specifications
- Execution logs
- Outcome analyses
- Capability assessments
- Updated MEMORY.md
- Updated learned-skills.md
- Updated skills
- Challenge log entries
- Scheduled challenge events
Requirements
- OpenClaw runtime
- sonnet4.5 model recommended
- Command execution access
- Read and write access
- Memory search access
- Sub-agent spawning access
