Versioned security assessment

Report ID: SA-EA3A0748

6/30/2026, 11:10:22 AM

ponyflash security assessment v2

Skill Security Certification Report

Audit History

Audit model: codex Historical report

Skill name: ponyflash
Version: v2
Maintainer: ponyflash
Source: github.com/ponyflash/ponyflash-skill/tree/main/
Coverage: 33 Files scanned · 5,103 Lines analyzed
Policy version: Unavailable

Highest confirmed finding severity

Medium

4 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 produced a high raw score because this community skill documents and ships shell scripts, FFmpeg workflows, network font downloads, and API-key setup. Review confirmed real medium-risk behavior that matches the stated media-generation purpose, while many Ruby backtick, weak-crypto, dynamic-import, and Windows SAM findings are false positives from markdown or comments. No prompt-injection attempt, credential exfiltration, hidden persistence, or confirmed malicious intent was found.

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

33 Files scanned · 5,103 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.

Observed in 3 evidence locations

Network access

May connect to external services.

Observed in 3 evidence locations

Filesystem access

May read or write local files.

Observed in 3 evidence locations

Env variables

May read values from the process environment.

Observed in 3 evidence locations

External commands

May invoke commands or programs outside the Skill.

Observed in 4 evidence locations

Risk findings

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

Confirmed security concerns (4)

RISK-001 Medium

scripts/media_ops.sh:176 scripts/media_ops.sh:226-231 scripts/media_ops.sh:285-290 scripts/media_ops.sh:401-402 scripts/media_ops.sh:613-616

Local FFmpeg Execution on User Media

The media workflow runs ffmpeg and ffprobe against user-selected input, subtitle, and output paths. This is expected for a media editing skill, but it can process untrusted media and write local files.

The script directly invokes ffmpeg and ffprobe for its documented media operations. Arguments are passed as shell words rather than eval strings, so command injection evidence is limited.

RISK-002 Medium

scripts/ensure_subtitle_fonts.sh:23-27 scripts/ensure_subtitle_fonts.sh:55-68 scripts/ensure_subtitle_fonts.sh:71-78 scripts/media_ops.sh:550-555

Network Font Download and Local Font Cache

The subtitle font helper downloads Noto Sans CJK SC from hardcoded public URLs, or from an environment-controlled URL. It writes the font into a cache or temporary font directory for subtitle rendering.

The network and filesystem behavior is explicit in the shell script and comments. The URLs serve a legitimate font purpose, but the environment override increases supply-chain risk.

RISK-003 Medium

SKILL.md:60-87 examples/quickstart.py:5-15 examples/quickstart.py:23-67

Cloud API Key and Account Calls

Cloud generation tasks require a PonyFlash API key and call PonyFlash account and generation APIs. This is core functionality, but users must understand that prompts, media, and credentials are used with a third-party service.

The skill clearly instructs users to provide or export an API key and then uses PonyFlash SDK methods. I did not find evidence that the key is sent anywhere except the intended SDK client.

RISK-004 Low

scripts/check_ffmpeg.sh:102-108 scripts/check_ffmpeg.sh:149-171

Dependency Check Prints Local PATH

The FFmpeg checker prints the operating system name and PATH during normal output. This can disclose local environment details in shared logs, but it does not write files or call the network.

The script prints platform, PATH, and tool paths when quiet mode is not set. This is useful diagnostics, but it is a real low-severity information disclosure concern.

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:70-87 SKILL.md:496-503 scripts/check_ffmpeg.sh:1-7 scripts/ensure_subtitle_fonts.sh:1-9

Static Analyzer False Positives in Documentation

Many reported Ruby backtick, dynamic import, weak-crypto, and Windows SAM hits are documentation or comment artifacts. Examples include fenced Python imports, markdown command examples, and the word SECURITY MANIFEST.

Targeted review showed these examples are markdown snippets or comments, not executed Ruby code or Windows SAM access. The same files contain real shell behavior, which is captured separately.

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