Back to skill

Security audit

Gmail Delete Email

Security checks for vulnerabilities and agentic risk

Overview

The skill does what it says, but it can move Gmail messages to trash from broad instructions without requiring explicit confirmation.

Review before installing. Only use this skill where the agent must ask you to confirm the exact Gmail message before trashing it, and ensure `gog` is authenticated only to the Gmail account you intend to manage.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill performs a destructive mailbox action but does not prominently warn that it moves mail to trash or require confirmation before doing so. In an autonomous or semi-autonomous agent workflow, this increases the risk of accidental data loss, especially when combined with the broad trigger condition and retry behavior.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The activation condition 'WHEN [An email needs to be deleted or moved to trash]' is broad and underspecified for a destructive action. In an agent setting, this can allow deletion based on weak inference, ambiguous user phrasing, or prompt-injected context rather than an explicit, confirmed user instruction tied to a specific message ID.

Content

No source excerpt is available for this finding.