Back to skill

Security audit

营销视频生成 Marketing Video

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed dLazy SaaS wrapper for ecommerce marketing video generation, with ordinary but real risks from installing a third-party CLI, uploading chosen files, storing an API key, and keeping project history.

Before installing, review the dLazy CLI source/package, prefer npx or an isolated non-privileged environment if you do not want a global binary, and attach only files you intend to upload to dLazy. Be aware that prompts, selected files, API credentials, and project history are handled by the dLazy service; use project clear/compact controls and rotate or revoke the API key if needed.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:5
Finding
Execution of Third-Party npm CLI Without Integrity Verification<![CDATA[ ## Vulnerability Details **File Location**: `SKILL.md:5, 61-64`; equivalent instructions in `SKILL-cn.md:5, 61-64` **Vulnerability Type**: Third-party supply-chain dependency executed without package integrity verification **Risk Level**: Medium ### Vulnerable Code From `SKILL.md`: ```yaml metadata: {"clawdbot":{"emoji":"🛒","requires":{"bins":["npm","npx"]},"install":"npm install -g @dlazy/cli@1.2.3","installAlternative":"npx @dlazy/cli@1.2.3","homepage":"https://github.com/dlazy-ai/cli","source":"https://github.com/dlazy-ai/cli","author":"dlazyai","license":"see-repo","npm":"https://www.npmjs.com/package/@dlazy/cli","configLocation":"~/.dlazy/config.json","apiEndpoints":["api.dlazy.com","files.dlazy.com"]},"openclaw":{"systemPrompt":"When this skill is called, run 'dlazy chat --skill product-to-ecommerce-video --prompt ...' for a new task, or 'dlazy chat --project <id> --prompt ...' to continue (discover ids via 'dlazy projects list'). Never pass both --skill and --project."}} ``` ```bash npx @dlazy/cli@1.2.3 <command> ``` ```bash npm install -g @dlazy/cli@1.2.3 ``` Equivalent instructions from `SKILL-cn.md`: ```yaml metadata: {"clawdbot":{"emoji":"🛒","requires":{"bins":["npm","npx"]},"install":"npm install -g @dlazy/cli@1.2.3","installAlternative":"npx @dlazy/cli@1.2.3","homepage":"https://github.com/dlazy-ai/cli","source":"https://github.com/dlazy-ai/cli","author":"dlazyai","license":"see-repo","npm":"https://www.npmjs.com/package/@dlazy/cli","configLocation":"~/.dlazy/config.json","apiEndpoints":["api.dlazy.com","files.dlazy.com"]},"openclaw":{"systemPrompt":"当调用此技能时,新任务运行 'dlazy chat --skill product-to-ecommerce-video --prompt ...',继续已有项目用 'dlazy chat --project <id> --prompt ...'(用 'dlazy projects list' 查 id);不要同时传 --skill 和 --project。"}} ``` ```bash npx @dlazy/cli@1.2.3 <command> ``` ```bash npm install -g @dlazy/cli@1.2.3 ``` ### Technical Analysis The skill instructs the agent or user to download and execute `@dlazy/cli` directly ...[truncated 2894 chars]
Remediation
<![CDATA[ ## Remediation Suggestions 1. Generate and maintain an approved lockfile containing the exact npm integrity digest for `@dlazy/cli@1.2.3`. 2. Verify the downloaded package against a trusted SHA-512 digest or signed release manifest before installation or execution. 3. Audit and vendor the exact reviewed CLI artifact when feasible instead of dynamically retrieving it through `npx`. 4. Establish reproducible-build or provenance verification, such as npm provenance attestations and a documented mapping between the reviewed source commit and published package. 5. Disable or strictly control npm lifecycle scripts where they are unnecessary. 6. Prefer an isolated, non-privileged execution environment with access only to explicitly attached files and required network destinations. 7. Do not run the installation with `sudo`, as an administrator, or under an account with broader permissions than the task requires. 8. Restrict outbound network access to the documented endpoints and monitor unexpected connections. 9. Protect `~/.dlazy/config.json` with owner-only permissions, use short-lived or narrowly scoped credentials where supported, and rotate the API key after any suspected package compromise. 10. Document a package-verification procedure in both language variants so users and agents do not execute an artifact based solely on its package name and version. ]]>
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The skill metadata advertises a broad marketing/promo video capability, but the actual execution path is hard-wired to `product-to-ecommerce-video`. This mismatch can cause the agent to invoke an unintended workflow, potentially sending user prompts and files to a narrower ecommerce-oriented remote service than the user expected, which is a real integrity and transparency issue.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The top-level branding says 'Marketing Video', but the operational instructions constrain all runs to an ecommerce product-video template. This deceptive or inaccurate labeling increases the risk of users sharing briefs, assets, or campaign materials under false assumptions about the tool's scope and destination, especially since attached files are uploaded to third-party infrastructure.

Session Persistence

Medium
Category
Rogue Agent
Content
---
name: dlazy-marketing-video
version: 1.0.2
description: "marketing video, promo video, ad video, brand video — create a marketing / promo video from a product, brand, or brief. Use when the user wants a marketing, promotional, or advertising video for social or campaigns."
metadata: {"clawdbot":{"emoji":"🛒","requires":{"bins":["npm","npx"]},"install":"npm install -g @dlazy/cli@1.2.3","installAlternative":"npx @dlazy/cli@1.2.3","homepage":"https://github.com/dlazy-ai/cli","source":"https://github.com/dlazy-ai/cli","author":"dlazyai","license":"see-repo","npm":"https://www.npmjs.com/package/@dlazy/cli","configLocation":"~/.dlazy/config.json","apiEndpoints":["api.dlazy.com","files.dlazy.com"]},"openclaw":{"systemPrompt":"When this skill is called, run 'dlazy chat --skill product-to-ecommerce-video --prompt ...' for a new task, or 'dlazy chat --project <id> --prompt ...' to continue (discover ids via 'dlazy projects list'). Never pass both --skill and --project."}}
---
Confidence
93% confidence
Finding
The skill explicitly instructs the agent to continue prior work via project IDs and states that chat sessions are tracked per project so follow-up turns keep context. This creates persistent remote state that may retain prior prompts, uploaded files, and generated content, increasing the risk of cross-task data exposure or unintended reuse of sensitive user information.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.