Back to skill

Security audit

create-technology-insight-report-rd

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed report-writing and QA workflow with helper checks, not a hidden or destructive package.

Before installing, make sure users understand that this skill can guide research using external sources and patent connectors. Do not send confidential R&D, patent, company, regulatory, or market material to external tools unless the user has explicitly authorized that use.

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
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (4)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared description promises a comprehensive report creation or rigorous review capability focused on substantive technology-insight content and evidence integration. The supplied code does not do that. It only validates an already-produced HTML report for structural and release-quality criteria such as required sections, metadata, version/date consistency, table widths, link portability, absence of scripts/stylesheets, placeholder text, legacy paths/domains, and legal disclaimer language. While this partially aligns with the phrase about auditing/localizing an existing report package, the actual functionality is much narrower and limited to QA checks on HTML packaging rather than substantive technology-insight review. Therefore the description materially overstates the skill’s primary purpose and capabilities.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The declared description promises a substantive report-generation or rigorous report-review capability covering evidence integration, analysis, auditing, and localization of a technology-insight report package. The supplied code only defines static checklist text and prints it to stdout via argparse. It performs no evidence collection, no report parsing, no HTML generation, no auditing logic, and no source-traceable analysis. While the checklist content is thematically related to technology-insight report QA, the actual primary behavior is much narrower: presenting SOP/checklist guidance. That is a materially different purpose from the declared end-to-end creation/review skill.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
83% confidence
Finding

The skill explicitly contemplates reading local files and using external sources/connectors, but it does not declare a bounded tool scope such as allowed tools, network limits, or data-handling restrictions. In a skill that processes confidential R&D, patent, regulatory, and company information, this omission can enable over-broad file access or unintended network disclosure if the hosting agent grants default capabilities.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The template sets <html lang="en">, which imposes English as the default language for generated reports. Under the natural-language policy rule, this is a locale choice without any visible opt-in or explanation that the template is intentionally limited to an English-only context.

Content

No source excerpt is available for this finding.