The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes external, third-party data from the Appllama catalog.
Ingestion points: Untrusted data such as app descriptions and screen names enter the agent context via tools like search_apps, list_app_screens, and get_screen (SKILL.md).
Boundary markers: The instructions do not provide explicit delimiters or warnings to ignore instructions that might be embedded in the retrieved metadata.
Capability inventory: The agent is empowered to write application code to the filesystem and execute shell commands for simulators and image generators (references/build-from-scratch.md).
Sanitization: There are no requirements for the agent to sanitize or validate the external content before using it to inform the build process.
[COMMAND_EXECUTION]: The instruction set requires the agent to utilize local system tools and simulators. Specifically, the playbooks guide the agent to run the iOS Simulator or Android emulator and execute image generation CLI tools to verify UI/UX implementations, which necessitates shell command execution.
[EXTERNAL_DOWNLOADS]: The skill facilitates downloading media assets (app screen images) from the Appllama service (appllama.io) into a local research/ directory for analysis and synthesis. These operations involve standard network and filesystem access to the vendor's infrastructure.

appllama-usage