ClawHub
Back to skill

Security audit

Token Cockpit

Security checks for vulnerabilities and agentic risk

Overview

Token Cockpit is a local token-cost analysis helper that reads OpenClaw usage logs, with no evidence of upload, destructive behavior, hidden persistence, or privilege escalation.

Install only if you are comfortable with a local tool reading OpenClaw usage logs. For ambiguous billing questions, explicitly tell the agent whether to analyze local logs and prefer passing a specific --logs path. Treat dollar figures as estimates unless you provide current pricing with --pricing.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
88% confidence
Finding
The skill instructs the agent to read local usage logs and inspect environment-based paths, which are capabilities with privacy implications, but it declares no permissions or trust boundary to the host. That mismatch can cause the agent to access local files or environment-derived locations without explicit user awareness or platform enforcement, increasing the risk of unintended data exposure.

Vague Triggers

Medium
Confidence
80% confidence
Finding
The invocation guidance is very broad and includes generic cost-related phrases, so the skill may activate in conversations about pricing or budgeting that are unrelated to local OpenClaw usage logs. In that context, the agent could inappropriately steer the interaction toward reading local files or making cost claims from incomplete data, creating unnecessary privacy exposure and confusing or unsafe tool use.

Static analysis

No suspicious patterns detected.