Skills nano-banana-edit

📦

nano-banana-edit

Name: nano-banana-edit
Author: runcomfy-com

Content revision r1 Medium Risk ⚙️ External commands🌐 Network access📁 Filesystem access

Edit Images with Nano Banana on RunComfy

Image edit workflows can drift when prompts do not protect identity, framing, or brand details. This skill gives Claude, Codex, and Claude Code a practical RunComfy workflow for targeted Nano Banana 2 edits.

Supports: Claude Codex Code(CC)

⚠️ 50 Poor

Install with my Agent

Copy this request to your Agent. It includes the canonical Skill page and manifest.

Agent request

Review the Skillstore skill "nano-banana-edit" from https://skillstore.io/skills/runcomfy-com-nano-banana-edit.md and its manifest at https://skillstore.io/api/skills/runcomfy-com-nano-banana-edit/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

Your Agent should still show its plan and request any confirmation required by the security policy.

Agent-readable resources

Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.

Markdown detail GET /skills/runcomfy-com-nano-banana-edit.md Signed manifest GET /api/skills/runcomfy-com-nano-banana-edit/manifest Signed lockfile GET /api/skills/runcomfy-com-nano-banana-edit/lockfile

Test it

Using "nano-banana-edit". A portrait needs a new studio background without changing the person or clothing.

Expected outcome:

The skill recommends a preservation-first prompt and a RunComfy Nano Banana Edit call using one public image URL.

Using "nano-banana-edit". A team wants ten product photos to share the same background and square framing.

Expected outcome:

The skill suggests a batch edit, locked aspect ratio, fixed resolution, and repeated prompt grammar across all inputs.

Using "nano-banana-edit". A label in one part of an image needs replacement while the rest stays unchanged.

Expected outcome:

The skill provides a spatially scoped instruction that names the exact area, retained elements, and replacement text.

Security Audit

Medium Risk

v4 • 7/5/2026 Open versioned report

Most static alerts are false positives caused by Markdown backticks, placeholder HTTPS URLs, or documented CLI credential storage. Confirmed issues are expected operational risks: the skill installs or runs the RunComfy CLI and sends prompts and image references to RunComfy infrastructure. No evidence found for prompt injection, arbitrary command injection, path traversal, or credential exfiltration.

Files scanned

186

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (1)

Medium

SKILL.md:177 SKILL.md:183

Third-Party Image and Prompt Processing

The skill sends edit prompts and image references to the RunComfy model API, and RunComfy fetches third-party image URLs for processing. This is expected behavior but can expose sensitive visual content to an external service.

The documentation explicitly states that the CLI posts to the RunComfy model API and that image URLs are fetched by the RunComfy model server.

Capability review items (9)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

Medium

SKILL.md:10

Ruby/shell backtick execution

Calls `runcomfy run google/nano-banana-2/edit` through the local

The skill states that it calls the local RunComfy CLI, which creates real command execution and network side effects. The command target is fixed, so injection risk appears limited.

Medium

SKILL.md:23-25

Ruby/shell backtick execution

```bash

The fenced example includes an npx command that installs this skill globally from an external source. This is expected setup guidance, but it still modifies the user environment.

Medium

SKILL.md:42-43

Ruby/shell backtick execution

1. **RunComfy CLI** — `npm i -g @runcomfy/cli`

The prerequisite tells users to install the RunComfy CLI globally with npm. That is legitimate for this skill but carries normal package supply-chain and host modification risk.

Medium

SKILL.md:43-44

Ruby/shell backtick execution

2. **RunComfy account** — `runcomfy login` opens a browser device-code flow.

The login command launches an authentication flow and can create local credentials for the RunComfy account. It is expected behavior, but it has credential-handling impact.

Medium

SKILL.md:67-74

Ruby/shell backtick execution

```bash

The example runs runcomfy with a JSON edit request and writes outputs to a local directory. The command is fixed, but it submits user-selected content to an external service.

Medium

SKILL.md:78-87

Ruby/shell backtick execution

```bash

The batch example runs the RunComfy CLI with multiple image URLs and output settings. This is intended functionality, but it is real external command execution.

Medium

SKILL.md:91-98

Ruby/shell backtick execution

```bash

The targeted edit example runs the RunComfy CLI against a user-provided image URL. The command is hardcoded, but it still performs local command execution and remote processing.

Medium

SKILL.md:177

Ruby/shell backtick execution

The skill invokes `runcomfy run google/nano-banana-2/edit` with a JSON body matching the schema. The

The text explicitly says the skill invokes runcomfy and the CLI posts to the RunComfy model API. This confirms expected but real command and network side effects.

Low

SKILL.md:177

Hardcoded URL

The skill invokes `runcomfy run google/nano-banana-2/edit` with a JSON body matching the schema. The

Line 177 documents a fixed RunComfy model API endpoint used for request submission. This is expected for the skill, but it means prompts and image references leave the local machine.

Audited by: codex View Audit History →

Share & cite this report

Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.

Open versioned report

Copy report link

https://skillstore.io/skills/runcomfy-com-nano-banana-edit/audits/4?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown badge

[![Skillstore security assessment](https://skillstore.io/badges/skills/runcomfy-com-nano-banana-edit/security.svg)](https://skillstore.io/skills/runcomfy-com-nano-banana-edit?utm_source=security_passport_badge)

HTML badge

<a href="https://skillstore.io/skills/runcomfy-com-nano-banana-edit?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/runcomfy-com-nano-banana-edit/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Embed card

<iframe src="https://skillstore.io/embed/skills/runcomfy-com-nano-banana-edit.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Academic citations (APA · BibTeX · CFF)

APA citation

runcomfy-com. (2026). nano-banana-edit security audit report (audit version 4) [Author version unspecified]. Skillstore. https://skillstore.io/skills/runcomfy-com-nano-banana-edit/audits/4

BibTeX citation

@techreport{runcomfy-com-runcomfy-com-nano-banana-edit-2026,
  author = {runcomfy-com},
  title = {nano-banana-edit security audit report (audit version 4)},
  institution = {Skillstore},
  year = {2026},
  number = {4},
  url = {https://skillstore.io/skills/runcomfy-com-nano-banana-edit/audits/4},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "nano-banana-edit security audit report (audit version 4)"
version: "unspecified"
type: report
authors:
  - name: "runcomfy-com"
date-released: "2026-07-05"
url: "https://skillstore.io/skills/runcomfy-com-nano-banana-edit/audits/4"
identifiers:
  - type: other
    value: "skillstore:runcomfy-com-nano-banana-edit:audit:4"
    description: "Skillstore immutable audit report identifier"

Compare variants

3 installable variants

Each author remains a separate installable skill. The recommended variant is ranked by Skillstore evidence.

Why this variant is first

Higher Skillstore usage

Author / source Skillstore Score Evidence Confidence Skillstore usage Updated

runcomfy-com Recommended Current

runcomfy-com-nano-banana-edit

Skillstore Score 50

Evidence Confidence High

Skillstore usage 10

Updated

2026-08-21

doany-ai

doany-ai-nano-banana-edit

Skillstore Score 50

Evidence Confidence High

Skillstore usage 8

Updated

2026-08-21

agentspace-so

agentspace-so-nano-banana-edit

Skillstore Score 38

Evidence Confidence High

Skillstore usage 6

Updated

2026-08-21

Skillstore Score

Why this score Evidence Confidence: High

Architecture

Maintainability

Content

Community

Spec Compliance

What You Can Build

Refresh Product Galleries

Create consistent background swaps for product images while preserving product shape, label details, and framing.

Adapt Campaign Creatives

Generate controlled ad variants with locked aspect ratio, stable subject identity, and repeatable prompt structure.

Localize Visual Assets

Replace a specific label, object, or background area while keeping the rest of the composition unchanged.

Try These Prompts

Simple Background Swap

Edit this image with Nano Banana Edit. Keep the subject identity, pose, clothing, and framing unchanged. Replace only the background with a clean studio setting.

Product Batch Consistency

Use Nano Banana Edit for these product images. Keep each product, label, angle, and scale unchanged. Apply the same neutral background and lock the aspect ratio.

Targeted Object Change

Edit only the object in the upper-right area. Keep the main subject, lighting, background, and all other objects exactly as shown.

Multi-Pass Brand Edit

Plan a two-pass Nano Banana Edit workflow. First preserve identity while changing the background. Then update the clothing color without changing the face or pose.

Best Practices

Lead every edit prompt with the details that must stay unchanged.
Use concrete spatial language for local edits, such as background only or upper-right corner.
Lock aspect ratio and resolution when editing a batch of related images.

Avoid

Do not combine many unrelated edit goals in one prompt.
Do not use private or sensitive image URLs unless RunComfy processing is acceptable.
Do not omit preservation goals when faces, products, logos, or labels must remain stable.

Frequently Asked Questions

What model does this skill use?

It uses the google/nano-banana-2/edit endpoint through the local RunComfy CLI.

Does it generate images from scratch?

No. This skill is for image-to-image editing from one or more existing image URLs.

How many input images can it handle?

The documented endpoint accepts 1 to 20 input images per call.

Can it keep a person or product unchanged?

Yes. The skill emphasizes preservation-first prompts for identity, pose, clothing, product shape, and framing.

Does it require a RunComfy account?

Yes. Users need the RunComfy CLI and either browser login or a RUNCOMFY_TOKEN value.

Where does processing happen?

The CLI submits requests to RunComfy model infrastructure and downloads generated outputs from RunComfy URLs.

Developer Details

Author

runcomfy-com

License

MIT

Skillstore revision

Version notice

The author did not declare a version.

Repository

https://github.com/runcomfy-com/skills/tree/main/nano-banana-edit/

Ref

62e2a730c5cd74eab4c7164309d810de660fcea3

Maintenance freshness

7/18/2026

Usage

5 downloads · 0 views

File structure

📄 SKILL.md

nano-banana-edit

Install with my Agent

Agent-readable resources

Test it

Security Audit

Confirmed security concerns (1)

Risk Factors

Copy report link

Markdown badge

HTML badge

Embed card

APA citation

BibTeX citation

CITATION.cff

Compare variants

Skillstore Score

What You Can Build

Refresh Product Galleries

Adapt Campaign Creatives

Localize Visual Assets

Try These Prompts

Best Practices

Avoid

Frequently Asked Questions

Developer Details