ClawHub
Back to skill

Security audit

survey-form-generator

Security checks for vulnerabilities and agentic risk

Overview

This survey generator is purpose-aligned, but it needs review because it embeds a service token and can send user survey goals to Dify's cloud API without explicit per-use confirmation.

Install only if you are comfortable sending the survey goal, audience, and intended use to Dify's cloud service. Avoid entering confidential customer lists, HR details, unreleased strategy, or sensitive business context unless you have approval. The generated HTML form appears local-only for responses, but the questionnaire generation step should be treated as a third-party API call.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Intent-Code Divergence

Low
Confidence
81% confidence
Finding
The README makes a strong privacy claim that historical conversation data and user identifiers are never sent to the remote service, but the document does not provide implementation details or enforceable guarantees that substantiate this. In an agent/skill context with broad invocation behavior, such claims can mislead users into disclosing sensitive business or customer information under false assumptions about what leaves the local environment.

Context-Inappropriate Capability

High
Confidence
99% confidence
Finding
The skill embeds a hard-coded Bearer token and instructs the agent to send user-provided survey goals to a third-party API. This creates two distinct risks: credential exposure/misuse and unintended exfiltration of user data to an external service not obvious from the skill’s high-level purpose. In skill context, this is more dangerous because the workflow is framed as a local survey generator/rendering tool, which can mislead operators about network use and data handling.

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The manifest and description present the skill as generating and rendering surveys, but the documented implementation sends the user’s prompt to a remote third-party service. That mismatch is a security-relevant transparency failure because users and reviewers may reasonably assume local processing when deciding what information to provide. The skill context increases risk because survey goals can contain sensitive business, HR, or customer research details.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The skill description instructs use for a very broad set of user intents, including loosely related phrases about understanding customer opinions, which creates ambiguous trigger boundaries. In agent environments, overbroad invocation can cause the skill to activate on conversations containing sensitive strategy, customer feedback, or internal planning data and route that content into a workflow or external service the user did not explicitly intend to use.

Static analysis

No suspicious patterns detected.