The Agent Skills Directory

[PROMPT_INJECTION]: The skill implements a persistent knowledge storage system by reading from and appending to references/learnings.md. This creates an indirect prompt injection surface where untrusted data or instructions from previous interactions could be stored and subsequently influence the agent's behavior in future sessions.
Ingestion points: The agent reads references/learnings.md during the context-gathering phase (Step 1) and processes user input.
Boundary markers: No delimiters or safety instructions are present to distinguish historical learned data from current system instructions.
Capability inventory: The skill instructions authorize the agent to read from and write (append) to the local filesystem.
Sanitization: No evidence of input validation, filtering, or sanitization is provided for the content being added to the knowledge file.
[COMMAND_EXECUTION]: The documentation provides an example command for installing additional platform components (npx skills add sales-skills/sales --skill sales-do). This involves executing a CLI utility to manage extensions within the platform's ecosystem.
[EXTERNAL_DOWNLOADS]: The suggested npx installation command involves downloading and executing packages from an external registry. This is documented as a standard procedure for users to acquire related tools from the vendor.

sales-momentum