Back to skill

Security audit

shop

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marked this skill benign because the reviewed material showed no clear malicious behavior, though it flagged privacy and workflow concerns and its analysis was incomplete.

The verdict is benign, not risk-free. Before installing, consider that image searches may transmit your image to Shop without a clear disclosure. The scan was partial: although all four listed components were scanned, some local references were unresolved, so the evidence does not cover those referenced paths.

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
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
80% confidence
Finding

Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Content

Scanner excerpt · references/catalog.md (reported line 36)May include surrounding context.

md
- Read the effective selection even when no warning appears; an unavailable combination can lose a requested option.

Image-similarity searches transmit image contents to Shop without a user-facing privacy disclosure.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The search instructions tell the agent to supply an image file's base64 bytes as search input. This sends the image contents to Shop, but the skill does not instruct the agent to disclose that transfer, which may expose private information contained in user-provided images.

Content

No source excerpt is available for this finding.

Image-search instructions omit a privacy warning about transmitting the supplied image.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The workflow instructs the agent to submit a local image for search, which transmits user-provided image data to an external service. The request-size discussion does not warn the user about this disclosure or advise against submitting sensitive photos.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/transactions.md (reported line 80)May include surrounding context.

md
## Read one order

Use `get_order` to reread an order or package from `search_orders`. Copy its `id` from `orders[]` or `shipments[]` unchanged. Sign-in is required; call it directly without checking the account first.