用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/hiyenwong/ai_collection --skill self-challenge命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Graph-native Python reimplementation of the Information Dynamics of Music (IDyOM) model that represents predictive memories as explicit graph objects for musical expectation modeling and network analysis.
Physics-aware end-to-end deep reinforcement learning methodology for quadcopter control with actuator dynamics modeling.
Reinforced Dreamer methodology for asymmetric reinforcement learning using latent guidance to improve world model representations and behaviors in model-based RL.
基于 SOC 职业分类
正在显示 SKILL.md
| name | self-challenge |
| description | Self-evolution skill that uses dual-agent challenge design and execution to expand capabilities over time. |
A self-evolution skill based on Agent0 paper (arXiv:2511.16043). Uses a dual-agent competition model where Curriculum Agent designs challenges and Executor Agent attempts them, driving continuous capability expansion.
开始自我挑战
挑战我的 [domain] 能力
本周自我挑战
The self-challenge mechanism uses dual-agent architecture:
┌─────────────────────────────────────────────────────────┐
│ Self-Challenge Architecture │
├─────────────────────────────────────────────────────────┤
│ │
│ ┌─────────────────┐ ┌─────────────────┐ │
│ │ Curriculum │ │ Executor │ │
│ │ Agent │──────▶│ Agent │ │
│ │ (设计挑战) │ │ (执行挑战) │ │
│ └─────────────────┘ └─────────────────┘ │
│ │ │ │
│ │ │ │
│ ▼ ▼ │
│ ┌─────────────────────────────────────────────┐ │
│ │ Review & Learn │ │
│ │ (回顾总结，提取经验) │ │
│ └─────────────────────────────────────────────┘ │
│ │ │
│ ▼ │
│ ┌─────────────────────────────────────────────┐ │
│ │ Knowledge Update │ │
│ │ (更新 skills, MEMORY.md, 工作流程) │ │
│ └─────────────────────────────────────────────┘ │
│ │
└─────────────────────────────────────────────────────────┘
Role: Design challenges that are slightly above current capabilities
Challenge Selection Criteria:
Challenge Types:
| Type | Description | Example |
|---|---|---|
| Tool Mastery | Learn to use a new tool | Use a new CLI tool to complete a task |
| Integration | Combine multiple tools | Create workflow combining 3+ tools |
| Problem Solving | Solve complex problems | Debug and fix a failing system |
| Knowledge Synthesis | Create new knowledge | Write a comprehensive guide |
| Automation | Automate repetitive tasks | Create a cron-based workflow |
Challenge Design Process:
Analyze Recent Performance
def analyze_capability_gaps():
# Review recent self-reflection reports
recent_reviews = read_recent_reviews()
# Identify weak areas
weak_areas = []
for review in recent_reviews:
if review.rating < 4:
weak_areas.append(review.area)
# Check for missing skills
existing_skills = list_skills()
recommended_skills = get_recommended_skills()
missing_skills = set(recommended_skills) - set(existing_skills)
return {
'weak_areas': weak_areas,
'missing_skills': missing_skills,
'improvement_opportunities': identify_opportunities()
}
Select Challenge Domain
def select_challenge_domain(gaps):
# Prioritize by impact
priorities = [
('weak_area', gaps.weak_areas),
('missing_skill', gaps.missing_skills),
('opportunity', gaps.improvement_opportunities)
]
for priority_type, items in priorities:
if items:
return {
'type': priority_type,
'domain': items[0],
'difficulty': 'medium'
}
Design Specific Challenge
## Challenge: [Name]
### Domain
[Capability area being tested]
[Easy/Medium/Hard]
[Clear, measurable goal]
Time limit: [duration]
Tools allowed: [list]
Success criteria: [measurable]
[What will be learned]
[How it improves capabilities]
[Available references]
[Similar challenges solved]
Role: Attempt to complete the designed challenge
Execution Process:
Understand Challenge
Execute
Document Results
## Challenge Execution Log
### Start Time
[timestamp]
### Approach
1. [Step 1]
2. [Step 2]
3. [Step 3]
### Obstacles Encountered
| Obstacle | Attempted Solution | Result |
|----------|-------------------|--------|
| [issue] | [solution] | [outcome] |
### End Time
[timestamp]
### Outcome
[Success/Partial Success/Failure]
### Key Learnings
- [Learning 1]
- [Learning 2]
Role: Analyze results and extract actionable knowledge
Review Process:
Outcome Analysis
def analyze_outcome(challenge, execution_log):
success_rate = calculate_success_rate(execution_log)
time_efficiency = calculate_time_efficiency(challenge, execution_log)
resource_efficiency = calculate_resource_efficiency(execution_log)
return {
'success_rate': success_rate,
'time_efficiency': time_efficiency,
'resource_efficiency': resource_efficiency,
'overall_score': (success_rate + time_efficiency + resource_efficiency) / 3
}
Capability Assessment
## Capability Assessment
### Before Challenge
| Capability | Level |
|------------|-------|
| [skill 1] | [1-5] |
| [skill 2] | [1-5] |
### After Challenge
| Capability | Level | Change |
|------------|-------|--------|
| [skill 1] | [1-5] | [↑/↓/=] |
| [skill 2] | [1-5] | [↑/↓/=] |
Knowledge Extraction
Update Knowledge Base
## Challenge: Master [Tool Name]
### Objective
Use [tool] to complete [specific task] with [success criteria]
### Constraints
- Time: 30 minutes
- Resources: Official docs, examples
- Success: Task completed without errors
### Steps
1. Read documentation
2. Set up environment
3. Execute basic commands
4. Complete target task
5. Document learnings
## Challenge: Integrate [Tools]
### Objective
Create a workflow that combines [tool1], [tool2], [tool3]
### Constraints
- Time: 60 minutes
- Must use all 3 tools
- Must solve real problem
- Success: Working workflow documented
### Steps
1. Identify integration points
2. Design workflow
3. Implement and test
4. Document and share
## Challenge: Solve [Problem]
### Objective
Debug and fix [failing system/error]
### Constraints
- Time: 45 minutes
- Must identify root cause
- Must implement fix
- Success: System working correctly
### Steps
1. Reproduce issue
2. Analyze logs/errors
3. Identify root cause
4. Implement fix
5. Verify solution
6. Document process
Recent self-reflection reports for identifying capability gaps.
Skills to be updated after challenges.
Long-term memory for storing challenge learnings.
Log of all challenges executed.
{
"challenges": [
{
"id": "challenge-001",
"date": "2026-03-05",
"domain": "tool-mastery",
"difficulty": "medium",
"outcome": "success",
"score": 0.85,
"learnings": ["..."]
}
]
}
openclaw cron add \
--name "self-challenge" \
--description "Weekly self-challenge for capability expansion" \
--cron "0 10 * * 1" \
--tz "Asia/Shanghai" \
--system-event "执行自我挑战任务"
If challenge completed in < 50% of allotted time:
1. Increase difficulty for next challenge
2. Add additional constraints
3. Expand scope
If challenge failed completely:
1. Break down into smaller challenges
2. Provide more resources
3. Adjust difficulty level
4. Reschedule with better preparation
If required resources are unavailable:
1. Find alternative resources
2. Modify challenge to use available resources
3. Postpone and acquire resources
User: "开始自我挑战"
Agent (Curriculum):
## Challenge: Master openclaw cron Commands
### Domain
Tool Mastery - OpenClaw CLI
### Difficulty
Medium
### Objective
Create, list, and manage 3 different cron jobs using openclaw CLI
### Constraints
- Time: 30 minutes
- Resources: openclaw docs, --help
- Success: 3 jobs created, listed, one executed
### Expected Learning
- Understanding of openclaw cron system
- Job scheduling best practices
- Error handling for cron jobs
---
Agent (Executor):
## Execution Log
### Start: 2026-03-05 11:00
### Approach:
1. Read openclaw cron --help
2. Create first job (daily)
3. Create second job (weekly)
4. Create third job (hourly)
5. List all jobs
6. Run one job immediately
### Obstacles:
| Obstacle | Solution | Result |
|----------|----------|--------|
| Unknown syntax | Used --help | Success |
| Timezone issue | Set tz explicitly | Success |
### End: 2026-03-05 11:25
### Outcome: Success (25 minutes)
### Key Learnings:
1. openclaw cron add syntax
2. Cron expression format
3. Timezone configuration importance
---
Agent (Review):
## Review Results
### Scores:
- Success Rate: 100%
- Time Efficiency: 83% (25/30 min)
- Overall: 92%
### Capability Change:
| Capability | Before | After |
|------------|--------|-------|
| openclaw cron | 2 | 4 ↑ |
### Knowledge Updates:
- Updated: learned-skills.md (new skill entry)
- Created: knowledge/topics/optimization/cron-jobs.md
### Next Challenge Suggestion:
Automate cron job monitoring and alerting
| Metric | Target | Measurement |
|---|---|---|
| Challenge Completion Rate | > 70% | Completed / Attempted |
| Capability Improvement | > 0.5/week | Average skill level change |
| Knowledge Creation | > 1/week | New skills or docs created |
| Time Efficiency | > 75% | Actual time / Allocated time |