The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/fetch_transcript.py executes sensitive system commands that modify network configuration.
Evidence: Uses subprocess.run to call wg-quick up wg0 and ip rule add from 10.100.0.2 table 51820.
Impact: These operations typically require root/sudo privileges and modify the system's global routing table and network interfaces.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through external data ingestion.
Ingestion points: Fetches video transcripts from YouTube and video metadata (titles/authors) from noembed.com in scripts/fetch_transcript.py.
Boundary markers: None identified. The skill directly returns raw text for the agent to process.
Capability inventory: The script contains subprocess.run calls for VPN management, although these are triggered based on VPN state rather than transcript content.
Sanitization: No sanitization or filtering is performed on the retrieved transcript text or video title before providing it to the agent.
[EXTERNAL_DOWNLOADS]: The skill requires external libraries and interacts with third-party services.
Fetches video metadata from noembed.com.
Requires installation of youtube-transcript-api and requests via pip.

youtube-transcript