reflect-check

What This Skill Does

The reflect-check skill serves as the primary diagnostic command for the SpecWeave reflection system within the OpenClaw ecosystem. Reflection is a critical process where the agent captures learnings, updates memory files, and refines future performance. When this system stalls, the agent loses its ability to learn from previous interactions. Reflect-check mitigates this by performing a six-point comprehensive audit: it validates the presence and structural integrity of reflect-config.json, scans the reflect.sh script for syntax errors, audits the last 10 execution logs, checks the memory file status to confirm growth, verifies the stop-reflect.sh hook, and runs a battery of pre-flight environment checks. By executing bash plugins/specweave/scripts/reflect-check.sh, the skill returns a color-coded status report, offering an immediate visual health readout. It transforms opaque, silent failures into actionable intelligence, allowing users to identify whether the issue is a broken configuration file, a missing dependency like jq, or a stalled automation hook.

Installation

To integrate this diagnostic tool into your OpenClaw environment, execute the following command via the CLI or terminal: clawhub install openclaw/skills/skills/anton-abyzov/sw-reflect-check Once the installation completes, the reflect-check commands become immediately available to the agent. No further environment variables are required, as the skill leverages existing SpecWeave file structures.

Use Cases

This skill is essential for maintaining high-performance agent behavior. Use it when you observe that the agent is not "learning" from mistakes or failing to update MEMORY.md files after complex tasks. It is also a mandatory step after performing marketplace updates, where script permissions or paths might have shifted. It is the go-to tool for developers debugging agent loops or silent failures, providing a clear audit trail of why the reflection automation might have stopped, been skipped, or failed to trigger.

Example Prompts

1. "Check if reflection is working; it feels like the agent isn't learning from our previous sessions."
2. "Run a reflect-check diagnostics scan to see why my memory files aren't updating."
3. "Reflection seems to be stuck, can you diagnose the issue?"

Tips & Limitations

• Tip: Run this skill immediately after modifying custom reflection logic to ensure syntax correctness.
• Tip: If you receive a 'Missing dependency' error in the report, ensure that jq is installed on your host system.
• Limitation: This tool is limited to the local SpecWeave reflection system. It cannot troubleshoot cloud-based memory synchronization issues.
• Limitation: The audit logs only track the most recent 10 attempts; for long-term historical analysis, consult the underlying system logs manually.