The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses railway-cli and executes bash scripts to perform project status checks and trigger deployments via a local library script (railway-api.sh).
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources (Railway project status and environment configurations) which serves as a potential ingestion point for untrusted content.
Ingestion points: railway status --json output and environment configuration GraphQL queries in SKILL.md.
Boundary markers: No explicit markers are defined to separate untrusted data from instructions.
Capability inventory: Shell execution through bash <<'SCRIPT' and Railway CLI operations.
Sanitization: No explicit sanitization or escaping of the ingested data is performed within the skill instructions, relying on the agent's parsing logic.
[DATA_EXPOSURE_EXFILTRATION]: The skill retrieves environment configuration but explicitly sets decryptVariables: false in its GraphQL queries, preventing the exposure of sensitive credentials to the agent's context.

railway-database