The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides tools to fetch data from external sources (web scraping, search results, document parsing). This data is untrusted and could contain malicious instructions designed to influence the behavior of an agent processing the output.
Ingestion points: SKILL.md defines tools for parsing documents via /augment/text-parser, scraping URLs via /augment/scrape, and performing web searches via /augment/search.
Boundary markers: No explicit boundary markers or instruction-delimiting guidelines are provided in the tool definitions to isolate the untrusted external content.
Capability inventory: The skill enables the agent to read local document files and fetch content from the public internet.
Sanitization: No sanitization or filtering logic is described; the tools return raw or markdown-formatted content directly from external sources.

venice-augment