ClawHub
Back to skill

Security audit

Accouting Skill For 财金记账

Security checks for vulnerabilities and agentic risk

Overview

This bookkeeping skill fits its stated purpose, but it can automatically create real financial records from inferred or OCR-parsed data without clear user approval.

Install only if you intend to use the caiwu888.cn accounting backend and can provide a scoped ACCOUNTING_API_TOKEN. Before using it on real books, require the agent to preview normalized fields and ask for approval before every add operation, especially for images, audio, PDFs, Excel files, or other parsed documents.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

High
Confidence
98% confidence
Finding
The skill instructs the agent to perform state-changing bookkeeping writes using inferred fields and defaults, explicitly without user confirmation. Because ledger, category, company, amount interpretation, and transaction type are AI-extracted and fuzzy-matched, a malformed prompt or ambiguous text can cause unauthorized or incorrect financial records to be created, corrupting books and triggering downstream business errors.

Missing User Warnings

High
Confidence
99% confidence
Finding
The multimedia workflow compounds the risk because OCR/STT/file parsing are error-prone and may misread amounts, dates, counterparties, or invoice semantics, yet the skill still proceeds to record entries without confirmation. An attacker or simply a noisy document can therefore induce incorrect financial writes from untrusted extracted content, making this more dangerous than plain-text entry.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill transmits potentially sensitive financial records to a third-party remote API, but the code provides no user-facing notice, consent checkpoint, or data-handling disclosure. In an agent context, users may assume local processing, so silent transmission of ledger, invoice, receivable, and payable data increases privacy and compliance risk.

Static analysis

No suspicious patterns detected.