ClawHub
Back to skill

Security audit

invoice-chasing-sequence

Security checks for vulnerabilities and agentic risk

Overview

The bundled skill keeps invoice follow-ups under your control, but its script produces only a basic calendar rather than the promised drafts and dispute guidance.

Expect a basic calendar from the script and review any agent-written drafts yourself. Keep invoice details private and use a dedicated output folder because plan.md is overwritten. The optional unsigned Windows app was not inspected and needs a separate trust decision.

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
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The code is loosely related to invoice collections through its title and guardrails, but its substantive output is only a dated schedule. The missing communication content and dispute handling are central to the declared purpose, making this a material functional mismatch. Local file reads and output-directory writes are supporting implementation details; no network activity, reminder sending, or other undeclared external actions are present.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_write.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The skill promises invoice follow-up drafts, call scripts, and dispute paths, but generates only a generic dated checklist (L083–L093).

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The manifest describes a 30/60/90-day invoice-chasing sequence with email drafts, a call script, and a dispute path for each touch. The implementation instead writes weekly dates labeled with the supplied focus or 'next block', followed by general guardrails; it generates none of the promised communication or dispute-handling content.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.