Openclaw Skill Self Improvement

• Outcome: Review
• Latest audit: May 29, 2026 at 9:35 PM UTC
• Version:
  1.0.0

Overview

The skill is a coherent local health-check tool, but it needs review because it reads local agent transcript logs outside the selected workspace and stores derived usage data.

Install only if you are comfortable with local Node scripts inspecting OpenClaw skill directories and agent transcript logs. Review generated .learnings reports before sharing them, and avoid running the daily heartbeat with untrusted workspace paths because it constructs shell commands from that path.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Context-Inappropriate Capability

Medium

• Confidence: 91% confidence
• Finding: The script recursively scans JSONL session transcript files from a hard-coded agents directory and uses their contents to infer skill usage, even though those transcripts may contain unrelated or sensitive agent interaction data. This creates an unnecessary data access surface and can expose private operational history without explicit user consent, especially because the path is outside the provided workspace root.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The script reads session transcript files and persists derived summaries without any disclosure, consent gate, or warning to the user. Because transcripts can contain sensitive prompts, identifiers, or operational details, silently processing them for a secondary purpose increases privacy risk and can surprise users who expect only skill metadata analysis.

VirusTotal

66/66 vendors flagged this skill as clean.