ClawHub
Back to skill

Security audit

OpenCode Zen - Check Free Models

Security checks across malware telemetry and agentic risk

Overview

This skill checks a public pricing page for free models and can update local model defaults only after asking the user.

Install only if you are comfortable with the skill making a curl request to the OpenCode Zen pricing page, creating ./zen_seen_models.json in the current workspace, and offering to change your local OpenClaw model defaults when it finds a new free model.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The skill says 'When this skill is triggered' but does not define clear invocation boundaries, while also instructing network access, state inspection, and potential configuration changes. In an agent environment, ambiguous trigger conditions can cause the skill to run in contexts the user did not specifically intend, increasing the chance of unsolicited external requests and follow-on prompts to modify local configuration.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to persist data to ./zen_seen_models.json automatically, but does not clearly disclose this storage behavior to the user before writing. Silent state writes create hidden persistence in the workspace, which can surprise users, interfere with repositories or automation, and normalize unauthorized local state changes by a user-invocable skill.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal