The Agent Skills Directory

[SAFE]: The skill instructions provide robust security guidance, requiring the use of encrypted credentials, tenant-bound authentication contexts, and HMAC signature verification for webhooks.
[INDIRECT_PROMPT_INJECTION]: The skill interacts with local repository files to analyze existing architecture.
Ingestion points: Uses Read, Glob, and Grep to inspect repository topology, schemas, and deployment files based on the [repository-path] argument.
Boundary markers: None explicitly defined for file contents, but the skill instructions provide specific mapping and definition rules to guide the agent's reasoning.
Capability inventory: Includes WebFetch for documentation and Write/Edit for producing architecture artifacts.
Sanitization: The skill relies on the agent's architectural modeling instructions to process discovered data into structured diagrams and models.
[EXTERNAL_DOWNLOADS]: The skill uses WebFetch to access Salesloft's official developer portal (developers.salesloft.com). These are trusted documentation resources essential for the skill's primary purpose.

salesloft-reference-architecture