← Back to Skills Marketplace
24
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install dreamcycle
Description
AI Agent Self-Reflection Engine — scan session logs, detect failure patterns, analyze recurrence trends, and suggest automated fixes.
README (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
Usage Guidance
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.
Capability Analysis
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.
Capability Assessment
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.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install dreamcycle - After installation, invoke the skill by name or use
/dreamcycle - Provide required inputs per the skill's parameter spec and get structured output
Version History
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.
Metadata
Frequently Asked Questions
What is DreamCycle?
AI Agent Self-Reflection Engine — scan session logs, detect failure patterns, analyze recurrence trends, and suggest automated fixes. It is an AI Agent Skill for Claude Code / OpenClaw, with 24 downloads so far.
How do I install DreamCycle?
Run "/install dreamcycle" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is DreamCycle free?
Yes, DreamCycle is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does DreamCycle support?
DreamCycle is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created DreamCycle?
It is built and maintained by qize-auto (@qize-auto); the current version is v0.1.0.
More Skills