solo-you2idea-extract

Warn

Audited by Gen Agent Trust Hub on Feb 21, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION] (MEDIUM): The skill relies on the Bash tool to execute a sequence of make targets and local CLI commands (uv run solograph-cli, make export, make deploy). These operations lack transparency and execute arbitrary logic on the host system, including processing external video URLs and channel names.- [EXTERNAL_DOWNLOADS] (LOW): Uses yt-dlp and uv to download external content (transcripts and packages) and relies on external search engines (Tavily/SearXNG), which introduces untrusted data into the local environment.- [DATA_EXFILTRATION] (LOW): Explicitly designed to egress local data to Cloudflare R2 and Pages. While this is the primary purpose, it creates a pathway for data exfiltration if the underlying make upload scripts are compromised or misconfigured.- [PROMPT_INJECTION] (LOW): Susceptible to indirect prompt injection (Category 8). The skill ingests YouTube transcripts into a database and analyzes them for insights. Ingestion point: VTT files via yt-dlp. Boundary markers: None. Capabilities: Bash, Write, Edit. Sanitization: None detected. Maliciously crafted subtitles could attempt to influence the agent's analysis or subsequent export steps.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 21, 2026, 05:54 PM