The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses Bash to execute commands involving solograph-cli and yt-dlp. It directly interpolates variables $URL and $CHANNEL, which are derived from user-supplied $ARGUMENTS, into shell command strings. This creates a surface for command injection if the input is not strictly validated by the agent.
[EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to install external packages solograph and yt-dlp using pip install. While yt-dlp is a common tool, solograph is a specific third-party dependency required for the skill's primary functionality.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from external sources.
Ingestion points: Video transcripts are downloaded via yt-dlp and saved to transcript.txt (SKILL.md).
Boundary markers: None. There are no instructions to use delimiters or ignore embedded commands within the processed transcripts.
Capability inventory: The agent has access to Bash, Write, Edit, and several solograph MCP tools, which could be abused if the transcript contains malicious instructions.
Sanitization: Absent. The skill uses sed to remove VTT formatting but does not filter or sanitize the actual text content for potential injection patterns.

solo-you2idea-extract