Versioned security assessment

Report ID: SA-4121DE96

8/6/2026, 10:40:18 AM

nano-banana security assessment v5

Skill Security Certification Report

Audit History
Scanner version 3.0.0 Audit model: codex Latest published report
Skill name
nano-banana
Version
v5
Maintainer
inferen-sh
Source
github.com/inferen-sh/skills/tree/main/tools/image/nano-banana/
Coverage
1 Files scanned · 148 Lines analyzed
Policy version
skillstore-security-audit-policy-v1

Highest confirmed finding severity

Medium

1 confirmed security finding requires attention.

Installation context

Check the current Skill page

This page summarizes report evidence only. The Skill page provides the canonical install advisory.

Open current Skill page

This report does not block or authorize the manifest or ZIP.

The skill legitimately invokes the belt CLI and remote Gemini image services, which can transmit prompts and source images externally. Most static findings are Markdown formatting, model identifiers, placeholder URLs, or documentation links without executable behavior.

Report position

Latest published report

Latest refers to the report sequence, not to artifact currentness.

Audit attestation

Active attestation

A public attestation is available for this exact report.

Human verification

Not verified

No human verification is recorded for this report.

Coverage

1 Files scanned · 148 Lines analyzed

11 items shown for review

Limitations

This report does not claim runtime or sandbox execution and does not prove the absence of side effects.

Evidence chain

Follow the evidence from source binding to the install contract. Available evidence supports verification; it is not a safety guarantee.

  1. Source

    Commit and path bound

  2. Artifact

    Content and tree hashes bound

  3. Audit

    Complete

  4. Install contract

    Open manifest to verify

    Open manifest

Capabilities observed

Observed means this report recorded supporting evidence. Not recorded does not prove that a capability is absent.

Contains scripts

May execute code included with the Skill.

Not recorded by this audit

Network access

May connect to external services.

Observed in 7 evidence locations

Filesystem access

May read or write local files.

Not recorded by this audit

Env variables

May read values from the process environment.

Not recorded by this audit

External commands

May invoke commands or programs outside the Skill.

Observed in 31 evidence locations

Capability review items (10)
Medium
SKILL.md:7
Ruby/shell backtick execution
> **Install the belt CLI skill:** `npx skills add belt-sh/cli`
Line 7 instructs users to run npx to install a remote skill. This executes third-party package tooling and introduces supply-chain risk.
Medium
SKILL.md:19-23
Ruby/shell backtick execution
```bash
Lines 19 through 23 direct the agent to authenticate with belt and run a remote image application. These are intentional external commands with account and service effects.
Medium
SKILL.md:43-47
Ruby/shell backtick execution
```bash
Lines 43 through 47 invoke belt to submit a text prompt to a remote Gemini image application. This has network, account, and possible billing effects.
Medium
SKILL.md:51-56
Ruby/shell backtick execution
```bash
Lines 51 through 56 invoke belt to request multiple generated images from a remote model. The command may consume service credits and transmit prompt content.
Medium
SKILL.md:60-65
Ruby/shell backtick execution
```bash
Lines 60 through 65 execute belt against a remote image model with a supplied prompt and aspect ratio. This creates external service effects.
Medium
SKILL.md:69-74
Ruby/shell backtick execution
```bash
Lines 69 through 74 invoke belt and pass an image URL to a remote editing model. The service can retrieve and process user-supplied image content.
Medium
SKILL.md:78-83
Ruby/shell backtick execution
```bash
Lines 78 through 83 run belt to request a 4K image from a remote model. The operation can transmit content and consume paid resources.
Medium
SKILL.md:87-92
Ruby/shell backtick execution
```bash
Lines 87 through 92 invoke belt with Google Search grounding enabled. This sends prompt content to remote services and may trigger additional data retrieval.
Medium
SKILL.md:118-125
Ruby/shell backtick execution
```bash
Lines 118 through 125 run belt, save a sample input file, and submit that file to a remote model. These commands affect local files and external services.
Medium
SKILL.md:129-138
Ruby/shell backtick execution
```bash
Lines 129 through 138 recommend three npx commands that install remote skills. Executing third-party installation tooling creates supply-chain and local modification risk.

Risk findings

Confirmed security concerns are separated from items that still need review.

Confirmed security concerns (1)

RISK-001 Medium
SKILL.md:43-46SKILL.md:69-73
User Content Sent to External Image Services
Image generation and editing commands send prompts and source image references to inference.sh and Gemini services. Sensitive content may leave the local environment.
The documented belt app run commands explicitly submit prompts and image URLs to named remote image models.

Remediation

Suggested fixes recorded by this audit. Applying them is the maintainer’s responsibility.

  1. FIX-001
    Medium
    Prompts and source images are sent to external inference services.
    Warn users before transmission and advise them not to submit confidential, regulated, or unlicensed content.
  2. FIX-002
    Medium
    The skill recommends remote skill installation through npx.
    Pin installation references to reviewed versions or immutable commits and require user confirmation before installation.
  3. FIX-003
    Low
    Generated images and grounded searches may consume paid service credits.
    Confirm model, image count, resolution, and search grounding before running billable commands.

Expert evidence

Immutable subject identity, scanner metadata, dismissed matches, and source-level evidence.

Artifact subject

Marketplace commit
4121de961d1b6f2ffca856260e239505c302452c
Content hash
492ac8d3e1667e0d083a0b01bd163195a707a6953f09250751f2fde0b26da8d9
Tree hash
37156bd66ad4487376d3760d5fe32e3418f5d468eda3f6412f35e1b3b60a87b0
Skill path
skills/inferen-sh/nano-banana
Audit payload hash
a7865a7f02917eee2d282e9e7797e500

Analysis metadata

Audit model: codex

Analysis state: Complete

Scope is limited to the recorded files, lines, methods, and evidence. No runtime or sandbox execution is claimed.

Verify and export

JSON report Signed attestation CITATION.cff SARIF 2.1.0 Signed installation manifest Source

The manifest and lockfile bind install artifacts to cryptographic hashes. This integrity claim is separate from the security assessment.

Audit attestation: active