The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses rg (ripgrep) to search for existing code patterns (pages, routes, components) and git diff --check for final validation. These are legitimate uses of development tools for context gathering and verification.
[PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes untrusted requirement documents provided by users.
Ingestion points: Reads external 'initial requirement documents' and repository AGENTS.md files (SKILL.md, Step 2).
Boundary markers: Absent; there are no instructions to use delimiters or ignore potential commands embedded in the requirement files.
Capability inventory: The skill has the ability to read and write files, as well as execute shell commands (rg, git).
Sanitization: No explicit sanitization or filtering of input data is mentioned, though it includes a human-in-the-loop step where it asks the user for confirmation before creating Pencil placeholder files.

feature-doc-splitter-cn