Security audit
Security checks for vulnerabilities and agentic risk
This is a coherent AssemblyAI transcription skill, with the main caveat that audio is sent to a third-party service for processing.
Before installing, assume any audio file or URL you transcribe will be processed by AssemblyAI. Do not use it for confidential, regulated, or highly sensitive recordings unless third-party processing is acceptable under your policies, and store the API key carefully if using the optional config file.
The README encourages users to submit audio for transcription via AssemblyAI but does not clearly warn that the audio content will be transmitted to an external third-party service. This can lead users to unknowingly upload sensitive conversations, voice messages, or meeting recordings, creating privacy, confidentiality, and possible regulatory exposure.
The skill declares that it requires an environment variable and, by its documented behavior, sends audio to AssemblyAI over the network, but it does not declare an explicit tool/permission scope. This can weaken least-privilege controls and reduce transparency for users and runners about what capabilities the skill needs.
The skill documentation describes transcription features but does not clearly warn users that submitted audio will be transmitted to a third-party service. This creates a privacy and data-handling risk, especially for meetings, interviews, or voice messages that may contain sensitive or regulated information.
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
## Setup
1. Create AssemblyAI account: https://www.assemblyai.com/
2. Get API key (free tier: 100 min/month)
3. Set environment variable:
The code is explicitly designed to communicate with an external API endpoint, which means audio data, URLs, and transcript-related metadata leave the local environment. In the context of a transcription skill this is functional rather than malicious, but it remains a real security/privacy concern if users are not clearly informed or if sensitive sources are processed.
Path(__file__).parent.parent.parent.parent / ".assemblyai_config.json",
]
BASE_URL = "https://api.assemblyai.com/v2"
def load_api_key():
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
return os.environ["ASSEMBLYAI_API_KEY"]
raise ValueError(
"No API key found. Create ~/.assemblyai_config.json with: "
'{"api_key": "YOUR_KEY"}'
)
The script uploads local audio content or forwards a user-supplied URL to AssemblyAI without any explicit consent prompt, privacy warning, or indication that potentially sensitive recordings will be sent to a third-party service. In a transcription skill, external transmission is expected, but the lack of disclosure increases the risk of users unintentionally exfiltrating confidential meeting, interview, or voice-message data.
The function emits fixed German strings such as '## Transkript', 'Sprache', and 'Dauer' regardless of user preference or locale. This is a natural-language locale policy concern because the skill imposes a specific language in output without opt-in or documented justification.
No suspicious patterns detected.