ClawHub

Tool结果大小控制器

Security checks across malware telemetry and agentic risk

Overview

The skill does what it claims: it saves oversized tool results to local files to reduce context size, with no network calls or credential access found.

Install only if you are comfortable with large tool outputs being written to local disk. Avoid using it on secrets, private records, or regulated data unless you adjust permissions, retention, and redaction yourself; run cleanup regularly and remember the preview may also expose sensitive content.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill explicitly describes automatically writing oversized tool results to local disk and returning a file path, but it does not warn that tool output may contain secrets, personal data, or other sensitive content that will now persist beyond the conversation. This increases exposure through local file access, backups, logs, and later reuse of spilled files, especially because previews and audit logs may also duplicate sensitive content or metadata.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The spill path writes oversized tool output to a predictable directory under the user's home folder automatically, with no consent gate, no sensitivity check, and no indication that confidential tool output may now persist on disk for up to 7 days. In an agent/skill context, tool results can contain secrets, prompts, API responses, or user data, so converting transient context into stored local files increases exposure to local compromise, accidental reuse, and privacy leakage.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal