ClawHub
Back to skill

Security audit

Clawlyra

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to do what it claims, but users should treat its gateway token as sensitive and remember that replies may be spoken aloud.

Install only if you are comfortable running a sidecar with an OpenClaw gateway token. Keep OPENCLAW_TOKEN in a private local environment file, exclude it from source control, rotate it if exposed, and use the skill where spoken responses will not disclose sensitive information to bystanders.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The README instructs users to place the OpenClaw gateway authentication token into Lyra's environment configuration but does not warn that this token is a sensitive secret requiring careful storage and restricted access. Because this skill relies on a sidecar app connecting as an operator, compromise or accidental disclosure of the token could allow unauthorized access to the user's OpenClaw gateway and agent capabilities.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly causes all model replies to be spoken aloud by a sidecar app, but the user-facing skill description does not clearly warn that sensitive content may be audible to bystanders. In a companion-style skill, users may share private or emotional information, so missing this warning increases the risk of unintended disclosure in shared environments.

Static analysis

No suspicious patterns detected.