Back to skill

Security audit

market-research-brief

Security checks for vulnerabilities and agentic risk

Overview

The bundled helper performs ordinary local file processing, but its output needs manual review and the optional Windows app requires separate assessment.

Use de-identified research and verify every numerical source manually. Expect a checklist rather than a completed brief, and choose an output directory where replacing market-research-brief.md is acceptable. The optional Windows executable is unnecessary for this workflow and is not covered by this verdict.

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)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The code shares the declared topic and predefined section names, but its primary output is a checklist of field lengths and missing values, not a market research brief. It neither transforms research notes into findings nor reproduces the submitted research content. Its limits use len(value), which counts characters, and its static guardrails do not enforce a source for every number. Local file reads and output writes are ordinary supporting implementation details; no unrelated resource access or hidden external activity is present.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_write.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The generated artifact is a field-validation checklist, not the sourced research brief promised by the manifest.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The manifest promises a research brief with fixed sections and a source for every number. The implementation reads field values but outputs only their presence and character counts, omitting the research content; L092 identifies the input file globally, and L116–L118 merely print sourcing reminders rather than associate sources with reported numbers.

Content

No source excerpt is available for this finding.

The skill enforces character limits instead of the advertised word limits.

Low
Category
Excessive Agency
Confidence
100% confidence
Finding

The manifest explicitly promises word limits, but the implementation compares len(value) against each limit and labels the results as characters. Consequently, content within the advertised word allowance can incorrectly be marked TOO LONG.

Content

No source excerpt is available for this finding.