The Agent Skills Directory

[PROMPT_INJECTION]: The skill operates by ingesting and synthesizing untrusted data, which establishes a surface for indirect prompt injection.
Ingestion points: The skill is designed to process external and untrusted data including customer transcripts, support tickets, and content from platforms like Reddit, G2, and YouTube (as specified in SKILL.md and references/source-guides.md).
Boundary markers: The instructions do not require the agent to implement boundary markers or specific delimiters when processing these inputs to differentiate between instructions and data.
Capability inventory: The skill leverages file read capabilities for local assets and web-scraping/search capabilities for online research gathering.
Sanitization: There are no defined requirements for the sanitization or escaping of external content before it is processed by the agent's context.

customer-research