Back to skill

Security audit

account-research-brief

Security checks for vulnerabilities and agentic risk

Overview

The bundled skill has limited local access, but its output requires manual source checks and its supplied template does not work with the script.

Treat the output as an unverified draft and check every claim, source URL, and date yourself. The script expects Markdown headings, despite the supplied YAML template. Choose an output directory where replacing digest.md is acceptable. Assess the optional unsigned Windows app separately; it is unnecessary for the bundled workflow.

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 (3)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The code provides a basic notes-to-digest formatter, which is related to the declared purpose but materially narrower. It copies matching section content verbatim into bullet points and accepts unsourced or undated claims. The input file path recorded in the output is not fact-level sourcing. No network retrieval, source verification, or date validation is implemented. No unrelated or covert capability is evident in the supplied chunk.

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 brief does not ensure that every fact has a source and a checked date, contrary to the manifest.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The script copies arbitrary note lines into the digest without requiring or validating a source URL and checked date; appending a guardrail does not enforce that requirement. Because the skill explicitly promises sourced, dated account research for live deals, unsupported, stale, or attacker-supplied claims could be mistaken for verified research and influence business decisions. This is an information-integrity weakness, not evidence of code execution or data exfiltration.

Content

No source excerpt is available for this finding.