The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze untrusted project files to determine deployment configurations.
Ingestion points: The agent reads local files including cmd/<app>/main.go, Dockerfile, and docker-compose.yml (as described in SKILL.md).
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potential commands embedded within the parsed source code.
Capability inventory: The skill possesses capabilities to create services, set environment variables, and manage deployments on the Railway platform.
Sanitization: There is no evidence of validation or sanitization for configuration keys or values extracted from the source files.
[COMMAND_EXECUTION]: The skill executes various shell commands via the Railway CLI (railway login, railway init, railway link, railway add) and system utilities (openssl, curl) to manage the deployment lifecycle and verify the health of the resulting services.
[EXTERNAL_DOWNLOADS]: The skill references the installation of the Railway CLI using standard package managers (NPM, Homebrew) which are well-known services for developer tooling.

genesis-railway