Security audit

Clawpify

Security checks across malware telemetry and agentic risk

Overview

This Shopify admin skill appears coherent and not malicious, but it gives broad live-store data and mutation guidance with incomplete safeguards for several sensitive operations.

Install only if you trust the operator and can use tightly scoped Shopify API credentials. Treat every write, delete, bulk export, webhook change, customer-data access, marketing-consent update, subscription change, fulfillment change, and customer notification as requiring explicit human review before execution.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (20)

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The file documents blog and article delete mutations without any caution that these operations permanently remove store content. In a high-privilege Shopify Admin context, an agent or user following this reference could delete published content unintentionally, causing integrity and availability harm to the storefront.

Missing User Warnings

Low

Confidence: 84% confidence
Finding: The create and update examples modify live blogs and articles, including publishing content, but do not warn that these actions can immediately change public storefront content. In a content-management skill, this omission increases the risk of accidental publication, defacement, or unintended edits by operators or autonomous agents.

Missing User Warnings

Medium

Confidence: 87% confidence
Finding: The documentation includes bulk export examples for orders and customers that expose potentially sensitive business and personal data, including customer identity, email address, purchase history, and order contents, without any warning about privacy, least-privilege scope use, or downstream handling of exported data. In a skill explicitly designed to query and manage Shopify stores, these examples can normalize broad data extraction and increase the chance of accidental over-collection or misuse by operators or downstream agents.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: This reference file provides ready-to-use GraphQL examples for listing, retrieving, creating, and updating customers, including personal data such as email addresses, phone numbers, physical addresses, order history, and spending information, but it gives no warning about sensitive-data handling, least-privilege access, consent, retention, or logging hygiene. In a Shopify Admin API skill focused on customer management, this omission can normalize broad access to PII and increase the risk of privacy violations, over-collection, or unsafe downstream use of customer data.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The fulfillmentEventCreate mutation modifies shipment history/status and can affect customer support workflows, internal logistics decisions, and the apparent state of delivery, yet this section provides no immediate confirmation warning. In a commerce operations skill, omitting a guardrail on shipment-status-changing actions increases the chance an agent executes a state-changing mutation without explicit user approval.

Missing User Warnings

Low

Confidence: 79% confidence
Finding: Releasing a fulfillment hold resumes order processing and can trigger downstream fulfillment actions, but the section itself does not warn that this is permission-gated. Although the later 'Dangerous Operations' summary mitigates risk somewhat, section-local omission can still lead an agent or integrator to treat it as a routine action and execute it prematurely.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: Moving a fulfillment order to a different location changes where inventory is sourced and where the order is processed, which can affect shipping timing, cost, stock allocation, and operational correctness. Because the section lacks a warning, an agent could perform a consequential logistics change without explicit consent or user awareness.

Missing User Warnings

Medium

Confidence: 87% confidence
Finding: Splitting a fulfillment order changes the fulfillment workflow by redistributing quantities across fulfillment orders, which can alter operational handling, partial shipment behavior, and downstream automation. Without an explicit warning at the point of use, an agent may perform this workflow-altering action without sufficient user review.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill documents two notification-sending mutations that trigger outbound contact to customers or recipients, but unlike the balance-changing and deactivation operations, it does not require explicit user confirmation before execution. This can lead to unauthorized or accidental email delivery, causing privacy issues, customer confusion, spam-like behavior, and reputational harm for the merchant.

Missing User Warnings

Medium

Confidence: 86% confidence
Finding: The skill documents a `locationDelete` mutation but does not prominently warn that it is a destructive, potentially irreversible administrative action. In a Shopify admin skill, users may copy and execute the mutation directly, leading to accidental deletion of fulfillment locations and operational disruption, especially because this skill is explicitly intended for live store management.

Missing User Warnings

Medium

Confidence: 79% confidence
Finding: The skill documents direct updates to email and SMS marketing consent states, including setting customers to SUBSCRIBED and SINGLE_OPT_IN, without any privacy, legal, or consent-validation warning. In a skill that enables operational API actions against real Shopify customer records, this can facilitate unlawful or unauthorized consent changes, exposing merchants to compliance violations and customer harm.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The documentation includes a destructive `menuDelete` operation and full-replacement `menuUpdate` behavior without any caution about confirmation, backup, or the risk of accidentally removing storefront navigation. In an agent skill context, this can encourage automated or unsafeguarded execution that disrupts a live store’s navigation and degrades availability or user experience.

Missing User Warnings

Medium

Confidence: 87% confidence
Finding: The documentation includes metafield deletion operations with no caution that the action is destructive and may permanently remove custom data from products or other resources. In a Shopify admin/integration context, users may copy these examples directly, so omission of warnings materially increases the risk of accidental data loss.

Missing User Warnings

High

Confidence: 95% confidence
Finding: The definition deletion example exposes a particularly dangerous operation because the mutation includes deleteAllAssociatedMetafields, yet the docs do not warn that deleting a definition can also remove all linked metafield data. This can cause broad, irreversible loss across many resources, making the omission more severe than a simple single-object delete.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The metaobject delete and bulk delete operations are documented without warning that removal may be irreversible and may affect structured content relied on by storefronts or admin workflows. Bulk deletion especially amplifies blast radius because multiple records can be removed in one operation based on provided IDs or conditions.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill explicitly requires confirmation for product publication and deletion, but bulk variant create/update operations also modify live catalog data and can materially change prices, SKUs, barcodes, and inventory-related attributes without any warning or approval flow. In a Shopify admin context, silent execution of these mutations can cause unintended storefront changes, pricing mistakes, fulfillment issues, or mass catalog corruption if triggered from ambiguous or malicious prompts.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The file documents deletion of Shopify delivery profiles without any warning that this changes live store shipping configuration and can affect checkout behavior, fulfillment coverage, and customer charges. In an agent skill context, omission of impact warnings increases the chance an automated or inattentive user triggers a destructive action against production store settings.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The create and update examples directly modify live delivery profiles, shipping zones, and rates but present them as routine examples without caution about production impact. In a Shopify admin skill, these actions can immediately change shipping prices, regional availability, and product-profile associations, making accidental misconfiguration more likely.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The file documents high-impact mutations such as pausing, activating, cancelling subscriptions, committing drafts, and skipping billing cycles without any user-facing cautions, approval requirements, or notes about business/customer consequences. In an agent skill that can directly operate against Shopify Admin APIs, this increases the chance an automated agent or user invokes destructive subscription actions unintentionally, causing billing disruption, customer impact, or irreversible contract state changes.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The webhook creation examples show how to send Shopify event data to HTTP, Pub/Sub, and EventBridge destinations, but they do not warn that these actions transmit potentially sensitive store, customer, order, or inventory data to external systems under the operator's control. In an agent skill context, omission of a warning or confirmation step can lead users to create outbound data flows without understanding the privacy, compliance, and data-exposure implications.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.