Security audit

awesome-demo-web-build

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent web project generator, but it asks agents to run unpinned remote scaffolding tools, fetch mutable remote Markdown, and ships an under-disclosed local Claude settings file.

Install only if you are comfortable with a skill that can scaffold a project, install npm packages, and fetch remote design/docs content. Before running generation, review the exact commands, prefer pinned package versions, avoid arbitrary or internal URLs, treat fetched DESIGN.md as untrusted design data, and check why the bundled Claude settings file disables figma-developer-mcp.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Insecure DependenciesIntroduces malicious components through unsafe dependency sources
Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
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

Findings (4)

T01 · Skill Instruction Hijacking

Warning

Location: SKILL.md:146
Finding: Mutable Remote Markdown Is Trusted as Agent Instructions
Content: View full analysis
Remediation: View remediation

T08 · Insecure Dependencies

Warning

Location: SKILL.md:138
Finding: Unpinned Package Runners Execute Mutable Third-Party Code
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: SKILL.md:140
Finding: Free-Form Values May Be Interpolated into Shell Commands Without Validation
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Note

Location: SKILL.md:91
Finding: Arbitrary User-Supplied URLs May Be Fetched Without Destination Restrictions
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection

Findings (37)

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The manifest describes a project generator but does not disclose that the skill may run scaffolding CLIs, install dependencies, fetch remote content, and write files into the project directory. This mismatch undermines informed consent and can cause users or orchestrators to invoke a much more powerful skill than they expect, increasing the risk of unsafe automated execution.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The activation scope is broad enough to match generic requests like 'build a demo' or 'create a web project', which can trigger the skill in contexts where the user did not intend command execution, dependency installation, or remote fetches. Overbroad routing increases the chance that risky actions are invoked unexpectedly under benign-seeming prompts.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The skill explicitly allows users to provide arbitrary reference URLs and instructs the agent to fetch them for design analysis. This broadens the skill from local project generation into unrestricted external network access, creating SSRF, data exfiltration, and retrieval of malicious content risks.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The skill directs the agent to run shell commands that scaffold a project, install packages, and modify the filesystem. This materially expands capability beyond merely providing project guidance: if auto-executed, it can perform arbitrary code execution through package scripts and cause persistent local changes.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 94% confidence
Finding: The skill executes npx create-next-app@latest, which pulls and runs remote code at execution time without pinning to a specific trusted version. This creates supply-chain and reproducibility risk: behavior can change unexpectedly, and a compromised upstream package or malicious release could execute arbitrary code on the agent host.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 95% confidence
Finding: The instruction npx shadcn@latest init fetches and runs the latest upstream package code at runtime. Because it is unpinned, the skill inherits supply-chain risk and non-deterministic behavior, which is especially dangerous in an agent workflow that may run commands automatically.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 95% confidence
Finding: The command npx shadcn@latest add [components] again executes remote code from an unpinned package, compounding the supply-chain risk. Since component names may vary by prior user selections, this also expands the range of code and files modified by the command in a way that is not strictly controlled.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 95% confidence
Finding: The template recommends executing npx create-next-app@latest, which fetches and runs the latest remote package version at execution time. In an agent skill that may be used to scaffold projects automatically, this creates a supply-chain risk because behavior can change unexpectedly or a compromised upstream release could be executed without review.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: The skill instructs use of npx shadcn@latest init, which executes code from an unpinned remote package. Because this is part of a reusable agent workflow, an upstream package compromise or breaking change could directly affect generated projects or the execution environment.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The document instructs users to include a third-party remote script from at.alicdn.com directly in the app without discussing integrity, privacy, CSP, or self-hosting tradeoffs. This creates a client-side supply-chain risk: whoever controls that script or the network path can affect every visiting user's browser.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 95% confidence
Finding: This blueprint includes npx create-next-app@latest in a project generation path, again relying on execution of the current upstream release. In the context of an agent skill, repeated encouragement of unpinned execution increases supply-chain exposure across every generated project.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: The npx shadcn@latest init command is unpinned and runs remote code during scaffolding. This is risky in an automated or semi-automated skill because it delegates trust to the package registry at run time.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: npx shadcn@latest add ... fetches executable package code without version pinning. Since the skill is meant to guide or automate project setup, this could introduce malicious or changed behavior into multiple downstream projects.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 95% confidence
Finding: This instance of npx create-next-app@latest has the same supply-chain weakness: it executes whatever upstream version is current when invoked. In a template library, this undermines reproducibility and increases the blast radius of any compromised release.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: npx shadcn@latest init pulls and executes the latest package code on demand. This creates avoidable risk for a build-generating skill, especially if end users may trust and run its commands without scrutiny.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: The command npx shadcn@latest add ... is another unpinned remote execution point. The cumulative pattern across the document suggests convenience was prioritized over integrity and reproducibility.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 86% confidence
Finding: The finding notes npx astro without a pinned version in the generic command table. Although the detailed example elsewhere uses npm create astro@latest, this command still encourages fetching and executing a moving upstream target, creating similar supply-chain risk.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 95% confidence
Finding: This npx create-next-app@latest command executes unreviewed latest code during SaaS app scaffolding. Because the skill may be repeatedly used to create production-like apps, a compromised release could affect many consumers quickly.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: npx shadcn@latest init is an unpinned code execution path. Within an agent skill, users may assume these are safe defaults and run them automatically, magnifying the risk.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: The unpinned npx shadcn@latest add ... command inherits the same supply-chain exposure. The skill context makes this more dangerous because it systematizes the insecure pattern across multiple templates.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 95% confidence
Finding: This content-site scaffold again uses npx create-next-app@latest, which is a moving trust boundary. If executed in CI, by an agent, or by users following the skill, it can introduce malicious or breaking behavior unexpectedly.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: npx shadcn@latest init executes the current release from the registry and is not reproducible. In a template repository, this is a real operational and supply-chain weakness rather than a harmless style issue.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: This npx shadcn@latest add ... command again trusts remote latest code at runtime. A malicious package update could alter generated files, introduce backdoors, or exfiltrate local environment data during execution.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 95% confidence
Finding: The tool template includes npx create-next-app@latest, another unpinned remote execution step. Given the skill's purpose is project generation, these commands are likely to be run, making the issue practical rather than theoretical.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 96% confidence
Finding: npx shadcn@latest init fetches and runs whatever the current release is. In agent-assisted workflows, this reduces user visibility into code provenance and increases exposure to supply-chain attacks.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.