Back to skill

Security audit

Audio Mastering CLI

Security checks for vulnerabilities and agentic risk

Overview

The skill appears intended for audio mastering, but it tells users to run a PowerShell script that is not included in the reviewed artifact and is obtained from an unpinned GitHub clone.

Review the exact scripts/master_media.ps1 content from the repository before installing or running this skill, prefer a pinned commit or signed release, and avoid running it in an elevated PowerShell session. Confirm input and output paths because the skill creates mastered media files on disk.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T03 · Remote Payload Retrieval and Execution

Error
Location
README.md:15
Finding
Unpinned Remote Repository Content Executed with PowerShell Policy Bypass<![CDATA[ ## Vulnerability Details **File Location**: `README.md:15-26`; `SKILL.md:16-18` **Vulnerability Type**: Remote payload retrieval and execution **Risk Level**: High ### Vulnerable Code From `README.md:15-26`: ```markdown ## Installation Clone this repo and place it inside your OpenClaw workspace `skills/` folder. ```powershell git clone https://github.com/alesys/openclaw-skill-audio-mastering-cli.git ``` ## Usage (examples) Audio to WAV + MP3: ```powershell powershell -ExecutionPolicy Bypass -File ".\scripts\master_media.ps1" -InputFile ".\water.wav" -MakeMp3 ``` ``` From `SKILL.md:16-18`: ```markdown ## Flujo 1. Verifica que existe el archivo de entrada. 2. Ejecuta: `powershell -ExecutionPolicy Bypass -File "{baseDir}/scripts/master_media.ps1" -InputFile "<ruta-archivo>" -MakeMp3` ``` ### Technical Analysis The installation instructions clone a mutable Git repository without pinning a reviewed commit, signed tag, release artifact, or cryptographic checksum. The usage instructions then execute `scripts/master_media.ps1` using PowerShell's `-ExecutionPolicy Bypass` option. The referenced script is absent from the audited artifact, which contains only `README.md` and `SKILL.md`. Its actual behavior—including command construction, filesystem access, network operations, FFmpeg invocation, and input validation—therefore cannot be verified. The effective executable payload is determined by the remote repository contents retrieved at installation time and may differ from the content that was originally reviewed. `-ExecutionPolicy Bypass` does not itself elevate operating-system privileges. However, it suppresses PowerShell execution-policy restrictions for that process, reducing a defense-in-depth barrier that might otherwise prevent or warn about script execution. This behavior best matches **T03: Remote Payload Retrieval and Execution**, because externally hosted, mutable content is retrieved and subsequently executed. ### Attack Path 1. A ...[truncated 1543 chars]
Remediation
<![CDATA[ ## Remediation Suggestions 1. Include `scripts/master_media.ps1` in the distributed and reviewed skill artifact so its behavior can be audited together with the documentation. 2. Pin installation instructions to an immutable, reviewed Git commit rather than cloning the mutable default branch: ```powershell git clone https://github.com/alesys/openclaw-skill-audio-mastering-cli.git Set-Location audio-mastering-cli git checkout --detach <reviewed-full-commit-hash> ``` 3. Prefer signed release artifacts and verify both the publisher signature and a documented SHA-256 checksum before execution. 4. Remove `-ExecutionPolicy Bypass` unless it is strictly necessary. If it is unavoidable, document the reason and constrain its use to the reviewed script and process. 5. Sign the PowerShell script and use an execution policy that validates trusted publisher signatures where supported. 6. Review the script for command injection, unsafe path handling, unintended network access, insecure temporary files, destructive output behavior, and improper quoting before distribution. 7. Execute media processing with least privilege, without administrator rights, and within a restricted working directory or sandbox where practical. 8. Add automated integrity checks that fail closed when the script hash or checked-out commit differs from the reviewed version. ]]>
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (5)

Hidden Instructions

High
Category
Prompt Injection
Content
---
name: audio-mastering-cli
description: CLI audio mastering without a reference track using ffmpeg; accepts audio or video inputs and outputs mastered WAV/MP3 or remuxed MP4.
metadata: {"openclaw":{"emoji":"🎚️","homepage":"https://github.com/alesys/openclaw-skill-audio-mastering-cli","os":["win32"],"requires":{"bins":["ffmpeg","powershell"]}}}
Confidence
60% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
---
name: audio-mastering-cli
description: CLI audio mastering without a reference track using ffmpeg; accepts audio or video inputs and outputs mastered WAV/MP3 or remuxed MP4.
metadata: {"openclaw":{"emoji":"🎚️","homepage":"https://github.com/alesys/openclaw-skill-audio-mastering-cli","os":["win32"],"requires":{"bins":["ffmpeg","powershell"]}}}
Confidence
60% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Natural-Language Policy Violations

Medium
Confidence
95% confidence
Finding
The primary invocation guidance is written in Spanish, which imposes a language choice on users without any opt-in or explanation that the skill is intended only for Spanish-speaking contexts. This matches the policy category for language or locale constraints that are not optional or justified.

Missing User Warnings

Low
Confidence
79% confidence
Finding
This is a markdown file, so SQP-2 applies to whether the skill description warns about behaviors affecting user data or system state. The usage section lists expected output files, but it does not explicitly disclose that executing the skill will create new files alongside the source media, which is a user-impacting filesystem action.

Missing User Warnings

Low
Confidence
84% confidence
Finding
This markdown file describes a workflow that generates mastered WAV/MP3/MP4 files, which is a user-data-affecting operation. Although outputs are listed, there is no explicit user warning or caution about file creation behavior, output location, or the need to review paths before execution.

Static analysis

No suspicious patterns detected.