Back to skill

Security audit

Neway Commerce OS

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent ecommerce storefront scaffolder with disclosed validation, build, mock payment, and mock AI pieces; it is not risk-free but the sensitive parts are expected for its purpose.

Before installing, understand that this skill creates and edits a generated project, runs npm commands, and includes placeholders for Stripe and an AI provider. Keep payments in mock/test mode until production-ready, store live secrets only in a secret manager or deployment environment, and review any generated storefront before deploying it publicly.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (52)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared description says this skill builds and verifies a complete ecommerce storefront skeleton from a short brief. However, the provided code does not generate, scaffold, build, or verify a storefront application at all. Its sole purpose is to scan repository files for Chinese/CJK text and fail if any is found outside narrow whitelisted locations. That is a materially different primary purpose and an unrelated capability compared with ecommerce site generation. Therefore this is a clear description-behavior mismatch.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The declared description promises an end-to-end skill that can generate, scaffold, build, and verify a commerce storefront from a short brief, implying substantive creation of site structure/content and execution of validation/build steps. The supplied code chunk is much narrower: it copies templates/base into a new directory, infers a site name/slug, replaces four placeholder tokens, backs up an existing target directory, and prints next-step instructions. It explicitly states it does not invent product data and relies on manual follow-up for products and pages. It also does not itself run validation or build commands. Therefore the actual behavior is materially narrower than the declared purpose, making this a clear description-behavior mismatch.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The declared description says the skill can generate, scaffold, build, and verify a commerce storefront skeleton from a brief. However, this code chunk does not generate or scaffold anything, does not transform a product brief into code, and does not run a build. Its primary function is enforcement of a predefined output contract for an existing project via static file/content checks. Verification is indeed part of the declared purpose, so the code is related to the skill domain, but the actual behavior of this chunk is materially narrower than the description and includes an undeclared optional network check for image URLs. Therefore the description does not accurately represent what this specific supplied code chunk actually does.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The declared description presents a broad code-generation/scaffolding skill for an entire ecommerce storefront. The actual code chunk only implements a narrow assistant API mock that returns hardcoded recommendation responses based on user-message keywords. This is a materially different primary purpose from generating or verifying a storefront skeleton. While an AI sales assistant is one feature mentioned in the description, this chunk alone does not perform the described end-to-end storefront creation behavior, so the description does not accurately represent what this code actually does.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The declared purpose describes a comprehensive ecommerce storefront scaffolding skill. The supplied code chunk only implements a placeholder Stripe webhook handler that inspects a request header and returns a static JSON response. This is a narrow backend integration scaffold, not evidence of generating or verifying a complete commerce storefront. While a Stripe webhook could be a supporting part of an ecommerce system, this code chunk by itself materially differs from the declared primary purpose and omits the core advertised capabilities.

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 76)May include surrounding context.

md
| T-04 | `lucide-react` in `package.json` dependencies |

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 265)May include surrounding context.

md
- The free-shipping threshold lives in `format.ts` as `FREE_SHIPPING_THRESHOLD`

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · templates/base/.gitignore (reported line 3)May include surrounding context.

text
node_modules
dist
.env
.DS_Store

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The README states that the repository self-check will error if the skill itself contains Chinese, which imposes a language constraint on maintainers or users of the skill. This is a natural-language policy issue because it forces a specific language policy rather than offering a language choice or clearly framing it as an optional mode.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The README instructs users to enable live Stripe mode and provide secret keys, but it does not include an explicit warning about handling production credentials securely or the risk of triggering real financial transactions. In a code-generation skill for ecommerce storefronts, this can lead users to paste real secrets into local files or test unfinished flows against live payment infrastructure, causing credential exposure or unintended charges.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding

The skill describes operations that involve reading and writing project files and optionally making network requests, but it declares no explicit tool scope or permission boundary. In an agent environment, missing scope declarations can allow the skill to be invoked with broader capabilities than users expect, increasing the risk of unintended file access, modification, or outbound requests.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

Because this is a markdown/manifest-style skill description, trigger specificity matters. The phrases describe a wide range of common storefront-generation requests but do not define clear limits, negative examples, or a narrow invocation context, which could lead to unintended activation for generic website-building requests.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The file contains natural-language instructions that force output language from environment detection rather than letting the user choose. This can violate language/locale policy expectations because the skill applies English-by-default and repository-English requirements without explicit user opt-in or a documented user choice mechanism.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

Line L65 states that the template ships with 'real copy in zh-CN' as a deliberate constraint. Because this is natural-language guidance in a markdown file and does not mention user selection, opt-in, or region-specific justification, it can be read as forcing a specific locale by default.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This markdown file describes redirecting users to Stripe Checkout and processing a webhook, which implies transmitting order and payment-related data to a third-party service. The section explains the mechanics but does not warn users about the external data transfer or privacy implications.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The local workflow instructs running the validator with --lang zh-CN --currency CNY, which imposes a specific language/locale and currency in natural-language guidance. The document does not indicate this is optional, user-selected, or justified as a region-specific deployment constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The docstring states that every file in the repository must be written in English and that everything else must be Chinese-free, with only narrow exceptions. This is a natural-language policy constraint that forces a specific language for a global audience without offering user choice or documenting a region-specific compliance justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The file's natural-language usage text states that --lang defaults to zh-CN, and the argument parser repeats that default later. This creates a built-in locale preference rather than requiring an explicit user choice, which is a language/locale policy concern for a general-purpose validator.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The parser sets --lang to default to zh-CN, so the script assumes a specific language unless the user overrides it. For a non-region-specific tool, forcing a language default without opt-in can violate organizational language/locale policy.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · templates/base/.env.example (reported line 9)May include surrounding context.

text
STRIPE_SECRET_KEY=
STRIPE_WEBHOOK_SECRET=

AI_API_BASE=https://api.deepseek.com/v1
AI_API_KEY=
AI_MODEL=deepseek-chat

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The assistant greeting and prompts are written only in English and instruct the user to interact in English, while the product data visibly includes multilingual fields such as nameEn and yen pricing. This can violate a language/locale policy if users are not given an explicit language choice or opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The helper explicitly formats numbers with toLocaleString('en-US'), which forces a specific locale for all users. The comments mention keeping the locale aligned with the target market, but the code does not offer any user choice or clearly document a justified region-specific restriction in this file.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The file states that keyword scenarios must match both Chinese and English keywords, which imposes a specific language policy in the skill behavior. There is no accompanying language-selection mechanism or explanation that this bilingual constraint is optional or context-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The argparse description exposes the same English-only enforcement in a user-facing string: the tool fails whenever Chinese text appears. That is a locale policy baked into the skill behavior rather than an optional or justified setting.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

This TypeScript file includes a natural-language/config value that forces a specific regional setting by always returning currency: 'CNY'. Under the policy, locale-related constraints should either be user-selectable or clearly documented as region-specific and justified.

Content

No source excerpt is available for this finding.