← 返回 Skills 市场
24
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install dreamcycle
功能描述
AI Agent Self-Reflection Engine — scan session logs, detect failure patterns, analyze recurrence trends, and suggest automated fixes.
使用说明 (SKILL.md)
DreamCycle
Detect, analyze, and auto-repair failure patterns from AI agent session logs.
When to use
Use DreamCycle when:
- An agent has been running for a while and you want to check for recurring failures
- Session logs are accumulating and you want trend analysis
- You suspect the same error keeps happening but no one is tracking it
- You want automated fix suggestions for known failure patterns
- After making changes to agent configuration, to verify nothing regressed
How it works
DreamCycle is a retrospective debugger for AI agents. It scans your agent's session log files in three phases:
Session Logs → Scan → Analyze → Fix
- Scan — Extracts failure, pattern, and lesson signals from JSON session files
- Analyze — Detects recurring patterns (appearing 2+ times) and trends (up/down/stable compared to previous scans)
- Fix — Maps known failure patterns to automated fix suggestions with confidence scores
Installation
pip install dreamcycle
Usage
# Basic diagnostic: scan last 2 days of sessions
dreamcycle diagnose ~/.openclaw/sessions/
# Scan last 7 days
dreamcycle diagnose ~/my-agent/logs/ --days 7
# JSON output for programmatic use
dreamcycle diagnose ./sessions/ --output json
# Version info
dreamcycle version
Example output
━━━ DreamCycle Diagnose ━━━
Scanned: 42 sessions (last 2 days)
Signals: 156 total (failures: 78, patterns: 52, lessons: 26)
━━━━━━━━━━━━━━━━━━━━━━━━━━
🔄 Top Recurring Patterns
[12x] "timeout" → ⚡ high
Fix: add retry_with_backoff (confidence: 85%)
[8x] "ModuleNotFoundError" → ⚡ high
Fix: pip install \x3Cmodule> (confidence: 95%)
📈 Trends
↑ timeout: 12x (up from 5x last scan)
↓ ModuleNotFoundError: 8x (down from 15x)
🔧 Auto-Fix Summary
✅ 2/4 patterns have automated fixes
❌ 2 patterns need manual review
Requirements
Only python3 and pip. Zero external dependencies — DreamCycle uses only
the Python standard library.
Notes
- Session files must be JSON format with a
messagesarray containingrole/contentfields - First run establishes a baseline; trends appear starting from the second run
- Trend data is stored at
~/.dreamcycle/scan_history.json
安全使用建议
This skill looks purpose-aligned and does not show malicious behavior in the provided artifacts. Before installing, verify the external dreamcycle package source and run diagnostics only on log folders you are comfortable processing, especially because agent session logs may contain private prompts, tool outputs, or secrets.
功能分析
Type: OpenClaw Skill
Name: dreamcycle
Version: 0.1.0
The DreamCycle skill bundle describes a diagnostic utility for analyzing AI agent session logs to identify failure patterns and suggest fixes. The documentation in SKILL.md outlines a standard workflow (Scan, Analyze, Fix) and specifies that the tool uses only the Python standard library, storing history locally in ~/.dreamcycle/. No malicious instructions, data exfiltration triggers, or obfuscated code were found in the provided files.
能力评估
Purpose & Capability
The stated purpose, examples, and requirements fit together as a retrospective debugger for agent session logs; the sensitive part is that those logs may contain private conversation or tool-output content.
Instruction Scope
The documented commands are user-directed diagnostics and version checks; the artifacts do not instruct hidden prompt overrides, automatic public posting, or automatic destructive repair actions.
Install Mechanism
The skill is instruction-only but tells the user to install a Python package with pip, so the runnable package code is outside the provided artifacts.
Credentials
Python and pip are proportionate for a CLI diagnostic tool, and no credentials or external network use are declared; however, the tool is meant to read local session-log directories.
Persistence & Privilege
The skill discloses local persistence at ~/.dreamcycle/scan_history.json for trend history; no elevated privileges or background persistence are shown.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install dreamcycle - 安装完成后，直接呼叫该 Skill 的名称或使用
/dreamcycle触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v0.1.0
Initial release of DreamCycle: AI Agent Self-Reflection Engine.
- Scans AI agent session logs to detect failure patterns and analyze recurrence trends.
- Suggests automated fixes for known failure patterns, including confidence scores.
- Supports command-line diagnostics with adjustable timeframes and JSON output.
- Stores historical scan data to report trends (increases/decreases in failure types).
- Requires only Python 3, with zero third-party dependencies.
元数据
常见问题
DreamCycle 是什么？
AI Agent Self-Reflection Engine — scan session logs, detect failure patterns, analyze recurrence trends, and suggest automated fixes. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 24 次。
如何安装 DreamCycle？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install dreamcycle」即可一键安装，无需额外配置。
DreamCycle 是免费的吗？
是的，DreamCycle 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
DreamCycle 支持哪些平台？
DreamCycle 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 DreamCycle？
由 qize-auto（@qize-auto）开发并维护，当前版本 v0.1.0。
推荐 Skills