The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from user-defined topics and external persona files.
Ingestion points: The skill ingests data via the <topic> argument and by reading external markdown files at paths such as ../[name]/reference.md and ../../framework.md using the Read tool.
Boundary markers: There are no explicit delimiters or instructions to the agent to ignore embedded commands within the ingested philosophical texts or user topics.
Capability inventory: The skill utilizes the Read tool and AskUserQuestion tool. It does not possess network, write, or shell execution capabilities.
Sanitization: No sanitization, validation, or escaping logic is present to process the external philosopher data or user-provided topics before they are incorporated into the agent's context.

dialogue