The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to process untrusted user queries regarding financial regulations and registration status, which serves as an ingestion point for potentially malicious instructions. This architectural pattern constitutes a surface for indirect prompt injection.\n
Ingestion points: User queries about investment adviser registration requirements, product classification, and regulatory boundaries as described in the 'When to Use' section of SKILL.md.\n
Boundary markers: The skill content lacks explicit delimiters or instructions for the agent to ignore embedded commands within user-provided text.\n
Capability inventory: The skill frontmatter authorizes the use of powerful system tools including Bash, Read, Write, and Edit.\n
Sanitization: There are no defined mechanisms or instructions for validating or sanitizing external user input before it is processed by the agent.

advice-standards