The Agent Skills Directory

EXTERNAL_DOWNLOADS (MEDIUM): The skill uses npm install -g @railway/cli to install the Railway CLI. While Railway is a reputable PaaS provider, it is not included in the predefined list of trusted organizations, making this an unverifiable external dependency.
COMMAND_EXECUTION (LOW): The skill heavily relies on the Bash tool to execute Railway CLI commands (railway up, railway run, etc.). These operations are necessary for the skill's primary purpose of deployment but represent a significant capability surface.
CREDENTIALS_UNSAFE (LOW): The skill instructions include patterns for reading sensitive data from .env.production and retrieving connection strings via railway variables get DATABASE_URL. While no hardcoded secrets are present, the skill is designed to handle and manipulate credentials, which requires careful agent handling to prevent accidental exposure.
INDIRECT_PROMPT_INJECTION (LOW): The skill possesses a vulnerability surface for indirect injection.
Ingestion points: Reads environment files (.env.production) and local repository content during deployment.
Boundary markers: None specified in the instructions.
Capability inventory: Full Bash access, network operations via CLI (railway up), and command execution within environments (railway run).
Sanitization: No evidence of sanitization or validation of the data being processed.

deploy-railway