🤖🤝🧠 better collab with your agent

• Analyzed: May 12, 2026, 3:19 AM
• Findings: 1
• Version: 1.0.0
• Hash: 3837b5c5b3a996cfa04e475ed927e7d75a0dc62bb54d989791dbebe629f6f9ba
• Source: palm
• External report: View on VirusTotal

Findings (1)

The skill `user-cognitive-profiles` is designed for local analysis of ChatGPT conversation exports to identify user cognitive archetypes. The `SKILL.md` and `README.md` provide clear, benign instructions for the user and do not contain any prompt injection attempts against the OpenClaw agent. The core scripts (`analyze_profile.py`, `compare_profiles.py`) perform local file I/O for input/output (e.g., `conversations.json`, `profile.json`) and do not exhibit any network communication, arbitrary shell execution, or access to sensitive system files. The `test_wildchat.py` script legitimately accesses the public 'allenai/WildChat-1M' dataset for testing purposes, which is not a security concern. All processing is local, aligning with the privacy claims in the documentation.