Back to skill

Security audit

Record screen, microphone or camera from macOS terminal

Security checks for vulnerabilities and agentic risk

Overview

This is a clear macOS recording skill, but it asks users to install an unpinned third-party recording executable that may receive microphone, camera, and screen permissions.

Review the publisher and Homebrew formula before installing, prefer a pinned reviewed release if available, and do not grant Microphone, Camera, or Screen Recording permissions until you intend to use a specific capture feature. Agents should ask before every recording, including screenshots and camera photos, and should avoid `--overwrite` unless the output path is explicitly confirmed.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:23
Finding

Unpinned Executable Installed from a Third-Party Homebrew Tap

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 23–27
Vulnerability Type: Untrusted and unpinned third-party dependency
Risk Level: Medium

Vulnerable Code

markdown
## Installation

```bash
brew install atacan/tap/record
text

### Technical Analysis

The Skill instructs users or agents to install an executable from the third-party Homebrew tap `atacan/tap` without specifying an immutable version, commit, checksum, or signature. Consequently, the installed implementation can change after this Skill has been reviewed.

This dependency is especially sensitive because the documented tool is intended to receive macOS Microphone, Camera, and Screen Recording permissions. The repository contains only usage documentation and no source code for the installed executable, so the audit cannot verify whether the external binary limits itself to the documented capture and local file-output behavior.

The installation instruction is not, by itself, evidence that the referenced package is malicious. The security issue is the absence of controls ensuring that future installations retrieve the same reviewed artifact.

### Attack Path

1. An attacker compromises the third-party tap, its formula repository, the formula’s upstream artifact source, or a maintainer account.
2. The attacker modifies the formula or referenced binary while retaining the expected package name.
3. A user or agent follows the documented `brew install atacan/tap/record` command.
4. Homebrew retrieves and installs the modified executable.
5. The user grants the executable one or more privacy-sensitive macOS permissions to perform the documented recording tasks.
6. The compromised executable abuses its execution context or granted capture permissions.

### Impact Assessment

Successful exploitation would execute attacker-controlled code with the privileges of the user running Homebrew and the installed command. Depending on permissi
...[truncated 498 chars]
Remediation
View remediation

Remediation Suggestions

  • Identify and link the authoritative source repository and the exact reviewed release of the recording tool.
  • Pin installation to an immutable release, formula revision, or source commit rather than implicitly accepting the latest tap contents.
  • Publish and verify a cryptographic checksum or trusted signature for the expected installation artifact.
  • Document the formula and upstream artifact sources so users can confirm that downloads originate from the intended publisher.
  • Prefer reproducible builds from reviewed source where practical.
  • Require users to verify the installed version and artifact identity before granting Microphone, Camera, or Screen Recording permissions.
  • Apply least privilege by granting only the macOS privacy permission needed for the requested capture operation and revoking it when no longer required.
  • Add a security note explaining that updating the package requires renewed review because updates can change executable behavior.
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (5)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill declares very broad trigger keywords such as 'record', 'camera', 'photo', and 'audio capture', which can cause the skill to activate in contexts where the user did not intend to invoke device recording capabilities. In this skill's context, accidental invocation is especially sensitive because the underlying commands can capture microphone, screen, or camera data, creating privacy and consent risks even though the document itself says consent should be obtained.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This reference documents commands for recording microphone and system audio, including mixed capture, but does not prominently warn that these features may capture private conversations, meetings, notifications, or copyrighted audio. In an agent skill context, normalizing such capture without explicit consent and privacy guidance can lead to unintended surveillance or unlawful recording, especially because the examples make capture look routine and easy to automate.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This reference documents camera and video capture functionality without any explicit privacy or consent warning, which can normalize or enable collection of sensitive visual data without informing users of the legal and ethical implications. In an agent skill context, concise command references may be reused directly by automation, making omission of consent guidance more dangerous because users or downstream systems may invoke recording in ways that surprise or surveil people.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The --audio option enables microphone capture, but the documentation does not warn that it records audio from the default microphone, which may collect highly sensitive conversations or bystander speech. In a skill or agent setting, undocumented audio capture increases the risk of covert recording because users may focus on camera use and overlook that sound is also being collected.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The --overwrite flag permits destructive replacement of existing files, but the reference lacks a warning that prior recordings or unrelated files at the target path may be permanently lost. In automation contexts, this can cause silent data loss if output paths are reused or generated incorrectly.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.