The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the web through integrated tools such as Web Search and Firecrawl. This external data is used to inform the generation of API code and specifications, presenting a potential surface for indirect prompt injection attacks.
Ingestion points: External content retrieved via Web Search and Firecrawl MCP servers during the analysis and design phases.
Boundary markers: No explicit instructions are provided to the agent to ignore or delimit instructions found in external content.
Capability inventory: The skill is designed to create a complex directory structure and write multiple file types including controllers, middleware, and configuration files.
Sanitization: No specific mechanisms for sanitizing or validating external input before interpolation into code generation tasks were identified.
[COMMAND_EXECUTION]: The workflow utilizes specialized platform commands (e.g., /sc:analyze, /sc:design, /sc:implement, /sc:test) to automate various stages of API development. These commands represent internal tool invocations within the intended execution environment.

api-builder