Security audit

tender-generator

Security checks across malware telemetry and agentic risk

Overview

This appears to be a legitimate bid-document generator, but it handles sensitive company and banking details in local Word and ZIP files.

Use this only in a trusted workspace. Treat the input JSON, templates, optional TMA.docx, generated Word files, and ZIP archive as confidential; prefer placeholder data for testing, choose a dedicated output directory, review files before sharing, and delete or secure outputs when finished.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection

Findings (8)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 89% confidence
Finding: The skill documents file-reading and local script execution behavior but does not declare corresponding permissions. Undeclared capability use weakens the platform trust model because users and reviewers cannot accurately assess what filesystem access the skill may exercise, especially when it processes uploaded templates and local files.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 96% confidence
Finding: The skill description promises a dialogue-driven generic document generator, but the documented behavior includes local JSON ingestion, hardcoded tender/bid document generation, modification of existing .docx files in an output directory, and ZIP packaging. This mismatch is dangerous because it can cause users to grant trust or provide sensitive material under false assumptions, while the actual automation may overwrite files or process data more broadly than expected.

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The trigger phrase "帮我生成一套项目文档" is broad enough that normal conversation about generating documents could unintentionally invoke this skill. In an agent environment, overly generic activation increases the chance of accidental execution, causing the skill to solicit sensitive company, pricing, or identity information and generate documents when the user did not explicitly intend to use this specific skill.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The markdown contains highly sensitive example data fields including legal identity, phone number, bank account, and banking routing details, without a strong handling warning or clear indication that all values are fictitious placeholders. In a document-generation skill centered on bids and corporate paperwork, this increases the chance that operators copy real production data into insecure contexts or that example values are mistaken for acceptable logging/test fixtures.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The usage guide embeds realistic-looking personal and financial data fields, including identity, phone, bank account, and banking routing details, without any warning to use synthetic placeholders or to protect sensitive data. In a document-generation skill specifically intended to process bid materials, this increases the chance that users will copy real regulated or confidential information into files, logs, examples, screenshots, or version control without adequate handling.

Missing User Warnings

Medium

Confidence: 85% confidence
Finding: The script persistently writes highly sensitive business data, including bank account details, company identity information, and contact data, into multiple .docx files and a ZIP archive without any explicit confirmation, minimization, or protection controls. In an agent-skill context, this increases the chance of accidental local disclosure, insecure retention, or unintended sharing of packaged documents containing sensitive information.

Missing User Warnings

Medium

Confidence: 86% confidence
Finding: The script generates multiple .docx files and a ZIP archive containing highly sensitive business and personal data, including company identifiers, contact details, and bank account information, and writes them to disk without consent prompts, overwrite safeguards, restrictive permissions, or data-minimization controls. In an agent skill context, this increases the chance of accidental disclosure through predictable output locations, unintended archiving, or silent replacement of existing files containing confidential bidding materials.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The README explicitly instructs users to prepare and place highly sensitive materials such as identity documents, business licenses, and audit reports into the workflow, but it provides no privacy, retention, access-control, or handling warning. In a document-generation skill, this omission increases the chance that operators will upload, store, or share regulated personal and financial data without appropriate safeguards, leading to privacy breaches or compliance violations.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.