Versioned security assessment

Report ID: SA-1F53481A

6/30/2026, 2:58:54 AM

higgsfield-soul-id security assessment v2

Skill Security Certification Report

Audit History

Audit model: codex Historical report

Skill name: higgsfield-soul-id
Version: v2
Maintainer: higgsfield-ai
Source: github.com/higgsfield-ai/skills/tree/main/higgsfield-soul-id/
Coverage: 3 Files scanned · 147 Lines analyzed
Policy version: Unavailable

Confirmed finding summary

No confirmed security findings

The completed audit recorded no confirmed security findings. This is not proof that the Skill has no side effects.

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 command and reconnaissance patterns. Most are false positives caused by documented Higgsfield CLI commands, but the remote installer is a true positive because it pipes a network script directly into a shell. No prompt injection text or confirmed malicious intent was found, so this is high risk rather than blocked.

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

3 Files scanned · 147 Lines analyzed

4 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.

Source

Binding unavailable
Artifact

Identity incomplete
Audit

Complete
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 1 evidence location

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 5 evidence locations

Capability review items (4)

High

SKILL.md:27-30

Remote Installer Piped To Shell

The bootstrap step downloads an install script from a GitHub raw URL and pipes it directly to sh. This is a true positive because the command executes remote code without pinning, signature verification, or a review step. It creates supply-chain risk even though the intent appears to be CLI installation rather than malware.

The exact curl-to-shell pattern is present and semantically used for installing an external CLI. I did not find evidence of malicious intent, but the execution risk is direct and well established.

Medium

SKILL.md:16 SKILL.md:31-32 SKILL.md:51-56 SKILL.md:63-71

Broad Bash Execution For External Service Workflow

The skill declares Bash access and instructs the assistant to run multiple Higgsfield CLI commands for account checks, training, waiting, generation, and listing IDs. This is mostly expected for the skill, but it still gives the skill a larger execution surface than a documentation-only workflow.

The commands are explicit and fit the stated purpose, which lowers concern. Risk remains because the skill relies on external command execution and account-state operations.

Medium

SKILL.md:44-54 references/photo-guide.md:7-14

Biometric Photo Upload Workflow

The workflow asks for 5 to 20 face photos and uses the CLI to upload local image paths or existing upload IDs. This appears necessary for Soul training, but it handles sensitive face data and should require clear user consent before submission.

The file requirements and upload behavior are explicit in the skill text. The privacy concern is semantic rather than malicious because the photos are central to the advertised capability.

Low

SKILL.md:51-56 SKILL.md:63-71 references/troubleshooting.md:20-24

Inline CLI Examples Triggered Command Detectors

Many external command alerts are examples of expected Higgsfield commands rather than Ruby backtick execution or dynamic shell construction. They are still relevant as a Bash-backed workflow but do not show command injection by themselves.

The commands are static examples with placeholders, not string-built shell execution. The analyzer label mentioning Ruby backticks does not match the Markdown content.

Risk findings

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

No confirmed security findings were recorded for this completed audit.

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:31-32 references/troubleshooting.md:5-24 references/photo-guide.md:31-37

Static Reconnaissance Alerts Are Mostly False Positives

The static reconnaissance alerts map to user-facing checks such as account status, paid plan status, training timeout guidance, and photo quality guidance. I did not find evidence of host enumeration, credential harvesting, or hidden system discovery in those lines.

The flagged text is visible workflow guidance and troubleshooting content. Confidence is high because the cited lines do not request system inventory or secret collection.

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