Back to skill

Security audit

Nano Banana 2 — Pro Pack on RunComfy

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward RunComfy image-generation helper with disclosed token use, network calls, and output handling.

Install only if you are comfortable sending prompts and generated-image requests to RunComfy, storing a RunComfy token locally or in RUNCOMFY_TOKEN, and paying any costs from higher resolutions, batches, or optional web-grounded generation.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • 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 (3)

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · SKILL.md (reported line 215)May include surrounding context.

md
- **Max 4 outputs per Nano Banana request.**
- **Nano Banana web search adds latency + cost** — only enable on demand.
- **Nano Banana 2K / 4K cost more** — default to 1K unless user asked for higher.
- **For Nano Banana image edit, use the dedicated edit skill** — not this endpoint.

## Exit codes

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest lists several concrete trigger phrases, but then expands scope to "any explicit ask to generate with Nano Banana," which is ambiguous and can overlap with requests intended for other Nano Banana variants or related image workflows. Because it does not define exclusions or negative examples, the activation boundary is unclear.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The FAQ states that English is the most reliable for prompts and that in-image text is best in Latin-script languages, which steers users toward a specific language/locale without presenting it as a user choice. This can conflict with organizational language policy when the skill does not explicitly offer multilingual use as an opt-in tradeoff.

Content

No source excerpt is available for this finding.