The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes external data from chat responses and search results.
Ingestion points: The output of the vs chat run command, which contains retrieved items and dialogue responses (SKILL.md).
Boundary markers: No specific delimiters or warnings to ignore embedded instructions are defined for the incoming search data.
Capability inventory: The skill has the capability to execute shell commands via the vs CLI (SKILL.md).
Sanitization: There is no mention of sanitization or validation of the external content before it is processed by the agent.
[COMMAND_EXECUTION]: The skill facilitates the execution of the vs CLI tool using subcommands such as chat run, app status, and app diagnose. These commands are used for interacting with the vendor's conversational search infrastructure.

vs-chat