The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The instruction file SKILL.md recommends a standalone installation method that downloads a shell script from a remote repository and pipes it directly into the bash shell for immediate execution (curl -fsSL https://raw.githubusercontent.com/yasserstudio/gpc/main/scripts/install.sh | bash). This pattern allows for the execution of unverified remote code without local inspection.
[COMMAND_EXECUTION]: The environment detection script scripts/detect_gpc.mjs utilizes execSync to run several shell commands to gather information about the environment, including gpc --version, npx gpc --version, gpc auth status, and gpc config get app.
[INDIRECT_PROMPT_INJECTION]: The skill processes external and user-provided data, creating a surface for potential indirect prompt injection attacks.
Ingestion points: The skill reads environment variables (such as GPC_SERVICE_ACCOUNT), configuration files (.gpcrc.json), and potentially receives data from the Google Play Developer API.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between its primary instructions and embedded data from these external sources.
Capability inventory: The skill's environment includes the ability to execute shell commands via the detection script and the gpc CLI itself.
Sanitization: No evidence of sanitization, validation, or escaping of the content from configuration files or environment variables was found before the data is utilized in command execution or detection logic.

gpc-setup