The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest user-supplied descriptions of features, data needs, and user actions to generate documentation, creating a surface for injection.
Ingestion points: User-provided text describing screen purposes, data requirements, user actions, and business rules during interaction with the agent.
Boundary markers: Absent. The instructions do not define delimiters or provide guidance to the agent on how to isolate user-provided content from the generation logic.
Capability inventory: The skill requires the agent to write files to the local file system (specifically in the .claude/docs/ai/ directory).
Sanitization: Absent. There are no instructions or mechanisms provided to validate or sanitize user input before it is written to the markdown documentation.
[COMMAND_EXECUTION]: The skill explicitly instructs the agent to perform file system operations based on user input.
Both SKILL.md and README.md mandate that 'ALL responses go to .claude/docs/ai//backend-requirements.md' instead of providing chat output. This persistent file modification capability, while central to the skill's purpose, allows for the creation of files based on untrusted user data.

frontend-to-backend-requirements