APIClaw

• Outcome: Review
• Latest audit: May 28, 2026, 9:09 AM
• Version:
  1.2.5

Overview

APIClaw is transparent about its proxy and credential model, but it gives an agent broad authority to call external APIs and may route full prompts or message content through a third-party proxy.

Install only if you are comfortable giving an agent a broad API gateway. Use dedicated least-privilege keys, avoid the proxy for sensitive prompts or message content, pin and audit the npm package, and require explicit human approval before sending messages, changing repositories, running paid jobs, or performing account-impacting actions.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: The skill advertises broad natural-language API discovery and execution capabilities without clearly constraining when direct API calls are allowed, what classes of APIs/actions are out of scope, or what user confirmation is required before side-effecting operations. In an agent setting, this ambiguity increases the risk of overbroad invocation, unintended external requests, and unsafe execution of high-impact actions such as code execution, scraping, messaging, or repository operations.

VirusTotal

65/65 vendors flagged this skill as clean.