ClawHub
Back to skill

Security audit

编码格式工具

Security checks for vulnerabilities and agentic risk

Overview

This skill is mostly a local encoding and format-conversion guide, but it adds unscoped API key, callback, and external API instructions that do not fit the stated local utility purpose.

Review this skill before installing. Its local encoding and file-conversion examples are ordinary, but avoid using callback_url, API keys, or remote API/curl flows unless you explicitly intend to send data outside your environment and know exactly what endpoint receives it. Do not process real secrets, JWTs, or private API responses through any remote callback path without separate confirmation.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Description-Behavior Mismatch

Medium
Confidence
88% confidence
Finding
The skill is presented as a local encoding/format conversion utility, but the documentation expands into external API usage, environment key setup, and callback behavior without a clear necessity tied to the stated purpose. This scope expansion can mislead an agent into handling sensitive data through networked workflows or preparing credentials where a local-only tool should not need them, increasing data exposure risk.

Description-Behavior Mismatch

Medium
Confidence
82% confidence
Finding
The examples introduce outbound HTTP usage via curl even though the skill's stated role is encoding/decoding and format inspection. In agent settings, such examples can normalize sending user-supplied data to remote endpoints, creating avoidable privacy and exfiltration risks if copied into execution flows.

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
The troubleshooting and operational guidance repeatedly references API authentication, token regeneration, network failures, and API calls despite those functions being outside the declared encoding tool scope. This broadens expected runtime behavior and may cause an agent or user to provision secrets or attempt network operations unnecessarily, increasing attack surface and confusion about trust boundaries.

Context-Inappropriate Capability

Medium
Confidence
91% confidence
Finding
The documentation introduces callback URLs and external API credential usage without a justified need for a local encoding/format conversion tool. Callback handling can create an implicit exfiltration path for processed data or metadata, especially if an agent interprets this as permission to POST results to arbitrary user-provided endpoints.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill describes callback/API behavior without a clear warning that data may be transmitted to external systems. In a tool expected to operate locally on potentially sensitive payloads such as JWTs, hashes, and encoded content, missing disclosure can cause unintentional leakage of secrets or private data.

Static analysis

No suspicious patterns detected.