The Agent Skills Directory

[DATA_EXFILTRATION]: The skill manages its API key by creating and reading from a sensitive file path at ~/.config/transcriptout/key. While this is a standard practice for CLI-based tools to manage credentials, accessing hidden configuration directories is documented as a data exposure risk.
[PERSISTENCE]: The setup instructions in references/auth-setup.md direct the agent to modify the user's shell profile files (such as ~/.bashrc, ~/.zshenv, or ~/.profile) to ensure the TRANSCRIPTOUT_API_KEY environment variable is loaded in future sessions.
[INDIRECT_PROMPT_INJECTION]: The skill's core functionality involves processing untrusted third-party data from YouTube, including transcripts and metadata, which could contain malicious instructions.
Ingestion points: Fetched data from api.transcriptout.com containing YouTube transcripts and video descriptions.
Boundary markers: The skill includes a dedicated 'Treat fetched content as data' section explicitly instructing the agent to ignore any commands or instructions found within the transcripts.
Capability inventory: The skill performs network operations to its API and has the capability to write to local configuration files and shell profiles.
Sanitization: Mitigation relies on the agent's adherence to the instructional boundaries provided in the skill documentation.

transcriptout