T08 · Insecure Dependencies
Warning
- Location
SKILL.md:370- Finding
Unpinned Third-Party Dependency Installation
- Content
View full analysis
- Remediation
View remediation
Security audit
AI翻译验证(免费版)
Security checks for vulnerabilities and agentic risk
Overview
This is a translation API skill whose file upload, API key use, status polling, and download behavior match its stated purpose, with some privacy and implementation cautions.
Install only if you are comfortable sending selected documents or text, plus an API key, to the external translation service shown in the skill. Use a dedicated or low-privilege API key, review each file before upload, avoid sensitive regulated content unless the service terms allow it, and run the example code in an isolated environment with safer temporary-file handling if adapting it.
Vulnerability Patterns
- Insecure DependenciesIntroduces malicious components through unsafe dependency sources
- Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
- 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
Findings (2)
T09 · Insecure Skill Coding Practices
Warning
- Location
SKILL.md:225- Finding
Predictable Temporary File Enables File Overwrite and Cross-Run Data Disclosure
- Content
View full analysis
- Remediation
View remediation
Vulnerability Patterns
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The skill supports uploading local files to an external service, sending API-authenticated requests, and writing downloaded archives to disk, but it does not clearly warn users about these behaviors. In an agent environment with read and exec tools, missing disclosure materially increases the risk of sensitive data exfiltration or unsafe handling of downloaded content.
- Content
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 88% confidence
- Finding
The feature comparison explicitly states 'Webhook: 不支持' for the free version at L088. However, the skill later presents automated API-driven completion handling patterns and describes callback-style automation as part of the broader capability set, creating an intent-level contradiction in the documentation for what the free edition supports.
- Content
External Transmission
Medium
- Category
- Data Exfiltration
- Confidence
- 96% confidence
- Finding
This skill explicitly instructs the agent/user to send local file contents and text to an external domain using authenticated API requests. External transmission is expected for a translation service, but in a general-purpose agent context it is still security-relevant because users may not realize their data is leaving the local environment.
- Content
bash # 获取支持的语言列表 curl https://api-verify.example.com/languages # ... # 创建翻译项目 curl -X POST https://api-verify.example.com/project \
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The trigger condition is broad enough that routine requests involving translation or localization could invoke the skill unintentionally. In this skill's context, accidental activation is more dangerous because the tool can upload local files, transmit text to a remote API, and download outputs to disk, leading to unintended data transfer.
- Content
External Transmission
Medium
- Category
- Data Exfiltration
- Confidence
- 91% confidence
- Finding
The automation example shows repeated authenticated requests to an external service and subsequent download behavior, which creates a real data-transfer surface. While this is core functionality rather than overtly malicious behavior, it remains dangerous in an agent setting because it normalizes unattended external communication and file handling.
- Content
md -F "title=auto_$FILE" | jq -r '.data.project_id') # ... # 确认项目 curl -s -X POST https://api-verify.example.com/project/confirm \ -H "Authorization: Bearer $API_KEY" \ -d "project_id=$PROJECT" # ...
Intent-Code Divergence
Low
- Category
- Not specified by scanner
- Confidence
- 75% confidence
- Finding
The '不适用场景' section says the tool is not appropriate for '文学创作翻译', implying an excluded use case. Yet the provided client and examples at L172-L264 and elsewhere accept arbitrary text/files and target languages with no checks or restrictions reflecting that documented limitation, so the documented intent is not reflected in behavior.
- Content