v1.0.1

Nano Banana Pro Custom

BenignClawScan verdict for this skill. Analyzed May 1, 2026, 8:00 AM.

Analysis

This image-generation skill appears purpose-aligned, but users should understand it sends prompts/images to a configured API provider using an API key.

GuidanceThis skill looks coherent for generating and editing images. Before installing, make sure the API base URL is a provider you trust, use a limited API key where possible, and avoid sending sensitive images or prompts unless you accept that provider’s data handling.

Findings (3)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Abnormal behavior control

Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.

Agentic Supply Chain Vulnerabilities

SeverityInfoConfidenceMediumStatusNote

scripts/generate_image.py

dependencies = ["openai>=1.0.0", "pillow>=10.0.0", "requests>=2.28.0"]

The script declares runtime Python dependencies with broad version ranges. This is normal for a uv-run Python tool but does not provide lockfile-level dependency pinning.

User impactFuture dependency versions could change behavior or introduce compatibility/security issues.

RecommendationInstall from trusted package indexes and consider pinning or reviewing dependency versions in controlled environments.

Permission boundary

Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.

Identity and Privilege Abuse

SeverityLowConfidenceHighStatusNote

SKILL.md

export NANO_API_KEY="your-api-key"

The skill requires and uses a provider API key, which is expected for an image-generation API integration but gives the skill delegated access to that provider account.

User impactUsing this skill may consume paid API credits or access models under the configured provider account.

RecommendationUse a dedicated, least-privileged API key if possible, monitor provider usage, and avoid sharing the configured key.

Sensitive data protection

Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.

Insecure Inter-Agent Communication

SeverityLowConfidenceHighStatusNote

SKILL.md

此技能兼容任何 OpenAI 风格的图像 API

The skill intentionally supports arbitrary OpenAI-compatible providers, so prompts and input images may be sent to whichever base URL is configured.

User impactPrompts and any input images may be processed by OpenAI, Azure OpenAI, OpenRouter, or another configured third-party provider.

RecommendationConfigure only trusted API base URLs and avoid sending sensitive images or private prompts to providers whose data-handling policies you do not trust.