ClawHub
Back to skill

Security audit

freelance-proposal-gen

Security checks across malware telemetry and agentic risk

Overview

The skill is an offline proposal generator, but its defaults can place a specific non-user WeChat/email contact and an on-by-default lead-generation watermark into client-facing proposals.

Before installing, review the contact defaults and either replace them with your own details or disable SHOW_CONTACT_WATERMARK. Generated proposals should be checked before sending because the skill is designed to add a visible lead-generation contact pill unless turned off.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
96% confidence
Finding
The script adds a lead-generation watermark derived from operator contact fields into every generated proposal by default, even though the skill is described as producing client-ready proposal documents offline. This is a hidden content injection behavior that can leak the creator's branding/contact details into user deliverables without clear user intent or manifest disclosure, creating integrity and privacy concerns.

Intent-Code Divergence

Low
Confidence
89% confidence
Finding
The module docstring claims the tool is a deterministic token renderer that substitutes caller-supplied fields, but the implementation also derives and inserts additional branding markup not directly present in the template input. That mismatch is security-relevant because it obscures behavior from auditors and users, undermining trust in generated business documents.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger phrases are broad, generic requests such as proposal-writing and quote-generation terms that can overlap with many normal productivity tasks. This can cause the skill to activate outside its intended niche, leading to misrouting, unexpected content generation, or bypass of more appropriate skills, especially in systems that auto-select skills from user text.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.