ClawHub
Back to skill

Security audit

Finance

Security checks for vulnerabilities and agentic risk

Overview

This finance skill is coherent, but it gives an agent under-scoped authority over sensitive accounting actions such as expense approval, payroll, payment, tax, and ledger operations.

Review this skill carefully before installing. It should be limited to drafting, validation, reconciliation support, reporting, and recommendations unless your environment enforces separate human approval for payments, reimbursements, payroll submission, tax filing, ledger changes, and period locks.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The phrase "When activated during a heartbeat cycle" assumes a triggering mechanism but does not define what a "heartbeat cycle" is, when it occurs, or what explicitly causes this skill to run. In a markdown skill description, that ambiguity can lead to unintended invocation because the trigger scope and exclusion conditions are not specified.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
88% confidence
Finding

The skill explicitly permits expenses under a threshold to be 'Auto-approved with receipt,' enabling the agent to participate in an approval decision without a mandatory human control. In a finance context, this weakens segregation of duties and could allow fraudulent, noncompliant, or duplicate reimbursements to pass if thresholds or receipt checks are manipulated or insufficient.

Content

Scanner excerpt · SKILL.md (reported line 208)May include surrounding context.

md
### Approval Workflow

Define thresholds:
- Under [Tier 1]: Auto-approved with receipt
- [Tier 1] – [Tier 2]: Manager approval
- [Tier 2] – [Tier 3]: Director/VP approval
- Over [Tier 3]: C-level/board approval