The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to perform project-wide searches using grep and file discovery via glob. These are used for legitimate purposes like finding hardcoded strings in source code and verifying audio asset paths.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The skill reads from src/ (source code), assets/data/strings/ (translation tables), design/gdd/ (game design documents), and design/narrative/ (narrative scripts).
Boundary markers: The instructions do not define specific delimiters or "ignore previous instructions" guards when reading external file content.
Capability inventory: The skill has access to Write, Bash, and Task (sub-agent spawning), which provides a significant action surface if an injection is successful.
Sanitization: No explicit sanitization or validation of the content read from source or translation files is mentioned before the agent processes the text.

localize