The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to run project-specific build commands such as npm run build or flutter build during its self-healing validation phase to ensure that copy changes do not break the application.
[DATA_EXPOSURE]: To perform its audit, the skill must read a wide variety of project files, including source code (JSX, TSX, Swift, Kotlin), configuration files (package.json, build.gradle), and documentation (README, style guides).
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes untrusted data from the codebase.
Ingestion points: Reads UI strings and documentation from source files like README.md, .tsx, .swift, .kt, and i18n JSON/YAML files.
Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions embedded within the strings being audited.
Capability inventory: The skill has the capability to write to the filesystem and execute arbitrary shell commands via build scripts.
Sanitization: There is no explicit sanitization or filtering of the content extracted from the project files before it is processed by the agent.

design-copy