Nano Banana Image T8

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:13 PM UTC
• Version:
  0.1.5

Overview

This skill is a disclosed Nano Banana image-generation helper that uses a user API key and saves outputs locally, with credential persistence that users should understand before use.

Install only if you intend to use the ai.t8star.cn Nano Banana image API through this skill. Treat the API key as a secret: prefer a dedicated limited-use key, know that it may be stored under ~/.whaleclaw/credentials/nano_banana_api_key.txt, delete or rotate it when no longer needed, and only provide image paths you are comfortable uploading to the API for editing.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 94% confidence
• Finding: The skill invokes bash, reads files, uses environment variables, persists credentials to disk, and calls a remote API, yet no explicit permission model is declared. That mismatch weakens operator visibility and policy enforcement, making it easier for a high-capability skill handling secrets to run with less scrutiny than it should.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill explicitly stores a user API key on disk for reuse but does not require a clear warning or consent flow about persistence, retention, and local exposure risk. Because the key is a reusable secret tied to an external service, silent persistence increases the chance of unintended credential storage and later compromise from local file access.

Ssd 3

Medium

• Confidence: 98% confidence
• Finding: The skill directs the agent to capture user-provided API keys from chat content and persist them for future sessions. Persisting secrets extracted from conversation text expands the blast radius of any prompt leak, transcript exposure, local compromise, or cross-session misuse, especially because the key enables authenticated outbound API access.

Ssd 3

Medium

• Confidence: 97% confidence
• Finding: These instructions operationalize secret extraction from user messages when certain trigger phrases appear, creating a rule-based mechanism for harvesting and saving 'sk-' style keys from free-form chat. Even with scope conditions, this is risky because conversational context can be ambiguous, logged, forwarded, or manipulated, and the stored key can be reused beyond the original intent.

VirusTotal

66/66 vendors flagged this skill as clean.