ClawHub
Back to skill

Security audit

Openclaw_Teach

Security checks across malware telemetry and agentic risk

Overview

The skill is purpose-aligned and disclosed, but users should review generated draft skills because it records local workflows and saves persistent instructions.

Install only if you are comfortable recording your screen locally. Avoid demonstrating secrets or private account details, decline optional browser-history access unless it is needed, and review the generated draft skill before relying on it or scheduling it.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding
The skill instructs the agent to invoke local Python scripts, read reference files, inspect generated media, optionally query a local DevTools endpoint, and write a new skill file under the user's home directory, but it declares no corresponding permissions. This creates a least-privilege and consent gap: a caller or reviewer may believe the skill is low-risk while it actually performs shell execution, file access, and persistent writes on the host.

Session Persistence

Medium
Category
Rogue Agent
Content
constants. If ambiguity would materially change the skill, ask the user
   concise questions and wait for answers.

7. **Write the skill.** Create
   `~/.openclaw/workspace/skills/<derived-slug>/SKILL.md` with:

   - frontmatter: `name` (lowercase-hyphen, ≤64 chars), `description` (one line,
Confidence
88% confidence
Finding
Write the skill.** Create `~/.openclaw

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.