Versioned security assessment

Report ID: SA-F849915A

6/28/2026, 6:14:40 AM

nano-banana-2 security assessment v2

Skill Security Certification Report

Audit History
Audit model: codex Historical report
Skill name
nano-banana-2
Version
v2
Maintainer
agentspace-so
Source
github.com/agentspace-so/runcomfy-agent-skills/tree/main/nano-banana-2/
Coverage
1 Files scanned · 207 Lines analyzed
Policy version
Unavailable

Highest confirmed finding severity

Medium

3 confirmed security findings require attention.

Installation context

Historical evidence

This report may not describe the currently installable artifact. Open the current Skill page for install guidance.

Open current Skill page

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

Static analysis reported many issues, but most backtick, weak-crypto, reconnaissance, and path-traversal alerts are Markdown documentation or CLI examples. Confirmed risks are legitimate external command execution, outbound RunComfy API use, optional web grounding, generated file downloads, and local token storage. No prompt injection attempt or malicious exfiltration intent was found, so this is publishable as a medium-risk community skill with disclosure.

Report position

Historical report

Open audit history before using this report to install.

Audit attestation

Not attestable

The required immutable binding is incomplete.

Human verification

Not verified

No human verification is recorded for this report.

Coverage

1 Files scanned · 207 Lines analyzed

3 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

    Binding unavailable

  2. Artifact

    Identity incomplete

  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 5 evidence locations

Filesystem access

May read or write local files.

Observed in 5 evidence locations

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 8 evidence locations

Risk findings

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

Confirmed security concerns (3)

RISK-001 Medium
SKILL.md:13SKILL.md:76-80SKILL.md:84-93SKILL.md:97-106SKILL.md:110-117SKILL.md:198
External RunComfy CLI Invocation
The skill instructs agents to run the local RunComfy CLI and pass user prompts to the google/nano-banana-2/text-to-image model. This is expected functionality, but it executes an external command and depends on correct argument handling by the caller.
The cited lines directly show CLI commands and the stated execution path. The risk is real but appears intentional for an image-generation skill rather than malicious.
RISK-002 Medium
SKILL.md:17SKILL.md:23SKILL.md:194SKILL.md:198SKILL.md:205
Third-Party Network Submission and Downloads
The skill sends prompts to the RunComfy model API and downloads generated results from RunComfy domains. This can expose prompt content and may fetch large generated files, although the endpoints are documented and related to the skill purpose.
The file explicitly names RunComfy web pages, the model API endpoint, and output download domains. The network behavior is confirmed and disclosed in the skill text.
RISK-003 Low
SKILL.md:79SKILL.md:92SKILL.md:105SKILL.md:116SKILL.md:202
Local Token File and Output Directory Handling
The skill documents local token storage under the user configuration directory and writes generated images to an output directory. This is standard CLI behavior, but users should understand that credentials and generated files touch the local filesystem.
The cited lines show output directory use and token file documentation. The behavior is sensitive but described as owner-only token storage, so severity remains moderate to low.

Expert evidence

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

Artifact subject

Marketplace commit
Unavailable
Content hash
Unavailable
Tree hash
Unavailable
Skill path
Unavailable
Audit payload hash
Unavailable

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.

Static false positives ignored (1)
Low
SKILL.md:4SKILL.md:9SKILL.md:25SKILL.md:37SKILL.md:46SKILL.md:60SKILL.md:70SKILL.md:132SKILL.md:183
Static Analyzer False Positives in Markdown Content
Several high and low static alerts are not supported by semantic review. The path traversal alert is a Markdown sibling-skill link, and the weak-crypto and reconnaissance alerts occur in descriptive text, tables, or headings rather than executable code.
Review of the cited lines found Markdown front matter, prose, tables, and a relative documentation link. No executable cryptographic logic, reconnaissance command, or filesystem traversal operation is present.

Verify and export

JSON report CITATION.cff SARIF 2.1.0 Signed installation manifest

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

Audit attestation: not_attestable