ClawHub
Back to skill

Security audit

Token Saver

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Claude cost-analysis helper that uses an Anthropic API key and local configuration review in ways that fit its purpose.

Before installing, be comfortable letting the skill read local OpenClaw prompt/config files and use your Anthropic API key to query Anthropic usage data. Review any proposed diffs before approving changes to openclaw.json or MEMORY.md, and consider using a least-privileged or rotated Anthropic key if available.

SkillSpector

Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill uses both environment access and network access (`ANTHROPIC_API_KEY` and a `curl` call to the Anthropic API), but it does not declare those capabilities as permissions. This creates a transparency and consent problem: users or the hosting platform may not realize the skill can read sensitive credentials and transmit data externally, increasing the risk of unauthorized or surprising data exposure.

Static analysis

No suspicious patterns detected.