ClawHub
Back to skill

Security audit

Accountsos

Security checks for vulnerabilities and agentic risk

Overview

This accounting skill is mostly coherent, but it should be reviewed because it encourages agents to create an external account and use read/write financial access without clear confirmation gates.

Install only if you intend to use AccountsOS for UK bookkeeping. Prefer creating the account and API key yourself, use the lowest scope that works, and require explicit approval before sending company or personal details, uploading documents, or creating or updating financial records.

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
  • 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 (4)

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill explicitly encourages an agent to self-sign up for an account and obtain a live API key, then immediately perform write-capable bookkeeping actions such as creating transactions, without requiring explicit human confirmation. In a finance context, this can cause unauthorized account creation, transmission of personal/company data, and modification of accounting records, which can create compliance, audit, and trust issues even without an attacker.

External Transmission

Medium
Category
Data Exfiltration
Content
**Option A: Self-Signup (recommended)** — Create an account yourself with one request:

```bash
curl -X POST https://accounts-os.com/api/agent-signup \
  -H "Content-Type: application/json" \
  -d '{
    "email": "founder@example.com",
Confidence
95% confidence
Finding
The self-signup flow instructs the agent to send user-identifying information such as email, company name, and full name to an external service without any built-in consent or verification step. Because the endpoint also returns a live read/write API key, this combines external data transmission with unauthorized account provisioning and creates a high-risk path for persistent third-party access to sensitive financial data.

External Transmission

Medium
Category
Data Exfiltration
Content
}

def call_tool(name, args={}):
    resp = requests.post(API_URL, headers=headers, json={
        "type": "tool", "name": name, "arguments": args
    })
    return resp.json()["result"]
Confidence
70% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Natural-Language Policy Violations

Low
Confidence
78% confidence
Finding
The description fixes the skill to UK limited-company bookkeeping behavior, and later sections reinforce UK-only assumptions. This can be a natural-language locale policy issue because the skill does not explicitly ask the user to confirm that a UK accounting context is desired before applying UK-specific rules.

Static analysis

No suspicious patterns detected.