The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from multiple untrusted external sources, which may contain malicious instructions designed to subvert agent behavior.
Ingestion points: The skill ingests content via web_fetch, extracts text from PDFs and documents, performs OCR on screenshots, and processes speaker-diarized transcripts from video/podcast URLs (SKILL.md).
Boundary markers: The skill does not specify the use of delimiters or instructions for the agent to ignore embedded commands within the ingested content.
Capability inventory: The skill possesses significant write capabilities, including put_page (database writes), add_link, add_timeline_entry, and gbrain files upload-raw for cloud storage operations.
Sanitization: There is no evidence of sanitization or validation of the fetched content before it is parsed for entity extraction and summarization.
[COMMAND_EXECUTION]: The skill utilizes shell commands to manage its knowledge base and handle file operations.
Evidence: The skill executes several subcommands of the gbrain CLI, including gbrain sync, gbrain files upload-raw, and gbrain files signed-url.
[EXTERNAL_DOWNLOADS]: The skill retrieves content and metadata from various external internet sources.
Evidence: Fetches raw content from user-provided URLs for articles, social media threads, and media files. It also interacts with cloud storage via the Supabase service for raw file preservation.

ingest