The Agent Skills Directory

[SAFE]: The skill follows a standard interactive workflow for project initialization.
[COMMAND_EXECUTION]: Executes basic shell commands (test -f and echo) to detect the project stack (e.g., Rails, Python, TypeScript). These operations are benign and intended for environment discovery.
[DATA_EXPOSURE]: Manages a local configuration file named compound-engineering.local.md. The skill does not access sensitive system files (like SSH keys or AWS credentials) or perform any network operations to exfiltrate data.
[PROMPT_INJECTION]: The instructions are clearly defined and do not contain attempts to override agent safety filters or bypass system instructions.
[INDIRECT_PROMPT_INJECTION]: While the skill reads an existing configuration file, it processes the information to either display it or guide the user through a reconfiguration process. There is no automated execution of instructions found within the data.

setup