ClawHub
Back to skill

Security audit

Product Business

Security checks for vulnerabilities and agentic risk

Overview

This is a broad product and business writing skill with some legal-template guidance, but it is disclosed, non-executable, and does not request system access or persistence.

Review the broad activation scope before installing, especially if you do not want one skill handling legal, sales, support, and product requests. Treat any legal output as a drafting aid only and have a qualified attorney review it before use.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Vague Triggers

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The metadata explicitly says to use the skill proactively across very broad categories like product planning, legal documents, support responses, and industry research. Over-broad proactive activation can cause the agent to seize control of many unrelated or only loosely related requests, leading to inappropriate instruction injection, mode misfires, and unauthorized generation of sensitive business or legal content.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The activation section spans eight major domains with extensive trigger lists and no strong disambiguation rules or refusal boundaries. This creates a large attack surface where common business terms can activate an overly powerful skill, increasing the chance of incorrect delegation, unexpected persona switching, and unsafe handling of regulated or sensitive topics.

Content

No source excerpt is available for this finding.

External Model or Provider Selection

Medium
Category
Excessive Agency
Confidence
90% confidence
Finding

Skill selects an external model or provider that may use a different account or billing plan than the operator expects. Undisclosed model switches can cause unexpected cost or quota consumption.

Content

Scanner excerpt · SKILL.md (reported line 10)May include surrounding context.

md
sales automation, customer support, or legal documentation.
  Use PROACTIVELY for product planning, business metrics, marketing campaigns,
  sales sequences, support responses, legal documents, or industry research.
model: sonnet
---

# Product & Business Skill

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The file states that each mode has its own language, and later specifies Product Manager as Chinese while many other modes are English. This can force a response language based on internal mode selection instead of user preference, which is a locale/language policy concern unless user opt-in is provided.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

These lines explicitly require Chinese for Mode 2A and English for English sections, which imposes a language policy tied to mode instead of the user's chosen language. There is no stated option to ask or adapt to the user's locale preference.

Content

No source excerpt is available for this finding.