Back to skill

Security audit

Work Productivity Nano Banana Workflow Helper

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill benign with high confidence: scans found no AIG alerts, while Skillspector noted broad activation terms that could cause it to be selected for unrelated requests.

The verdict is benign, but broad activation terms could lead to unintended skill selection. Consider whether that would be acceptable for your use; the reported scan covered all six components, with no partial or uninspected files.

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
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Activation keywords include broad everyday terms without domain-specific constraints.

Medium
Category
Trigger Abuse
Confidence
97% confidence
Finding

The keyword list includes standalone terms such as pro, generate, edit, and images, which overlap with many requests unrelated to Nano Banana Pro workflows. No exclusion conditions or requirement to combine these terms with a specific product or workflow context is provided.

Content

No source excerpt is available for this finding.

Activation keywords include common, unscoped terms that can trigger the skill for unrelated requests.

Medium
Category
Trigger Abuse
Confidence
99% confidence
Finding

The trigger list includes everyday terms such as banana, pro, generate, and edit without requiring a Nano Banana workflow context. The description at L04 also permits activation on standalone generic terms, and no exclusions are provided, making unrelated food, editing, or general productivity requests potential matches.

Content

No source excerpt is available for this finding.

Standalone trigger keywords are too broad for a Nano Banana workflow skill.

Medium
Category
Trigger Abuse
Confidence
99% confidence
Finding

Keywords such as generate, edit, images, pro, and banana occur in many unrelated requests. The skill does not require these keywords to appear in a Nano Banana workflow context or provide exclusion conditions, creating a clear risk of unintended activation.

Content

No source excerpt is available for this finding.

The invocation description does not establish a clear activation scope despite allowing implicit invocation.

Medium
Category
Trigger Abuse
Confidence
90% confidence
Finding

The short description and default prompt contain fragmented demand commentary rather than a clear definition of supported tasks or activation conditions. Although they mention Nano Banana Pro workflows, they do not clarify which requests qualify; with implicit invocation enabled, this leaves skill selection ambiguous.

Content

No source excerpt is available for this finding.

Activation keywords include broad everyday terms without explicit scope constraints.

Low
Category
Trigger Abuse
Confidence
97% confidence
Finding

Keywords such as pro, generate, edit, and banana overlap with ordinary requests unrelated to Nano Banana Pro workflows. The usage section does not require these keywords to appear in a Nano Banana-specific context or provide exclusions, creating a risk of unintended invocation.

Content

No source excerpt is available for this finding.

Usage keywords include broad everyday terms without a domain-scoping requirement.

Low
Category
Trigger Abuse
Confidence
91% confidence
Finding

The usage signals include standalone terms such as "generate", "edit", "images", "pro", and "bug fix", which overlap with many unrelated requests. Although the example trigger sentences mention Nano Banana Pro, the file does not require that context when matching keywords or provide exclusions.

Content

No source excerpt is available for this finding.