The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to set up automated persistence and background tasks using system utilities like cron, systemd, or launchd when an automated synchronization workflow is requested.
[EXTERNAL_DOWNLOADS]: The skill requires several external packages and CLI tools for content ingestion, specifically recommending installations via npm (such as agent-browser) and pip (such as yt-dlp and pdfplumber), and relies on specialized tools like defuddle and xurl for data extraction.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because its core functionality involves the ingestion and synthesis of unvalidated data from the web.
Ingestion points: Data is pulled from Twitter threads, YouTube transcripts, and various web URLs into a raw/ staging area before being processed into the knowledge base.
Boundary markers: The instructions lack specific guidance on using delimiters or protective prompts to prevent the agent from executing instructions potentially hidden within the fetched content.
Capability inventory: The agent possesses the capability to execute shell commands, perform network requests, and modify files across the local filesystem.
Sanitization: The skill does not prescribe any validation or sanitization procedures for ingested content before it is incorporated into the canonical knowledge base.

second-brain