The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the Railway CLI using curl | sh and bash process substitution from railway.com and agents.railway.com. These are official domains owned by the skill's author and are used for legitimate tool installation.
[COMMAND_EXECUTION]: Multiple Python and Bash scripts (dal.py, analyze-postgres.py, enable-pg-stats.py, etc.) execute Railway CLI commands and database client binaries (psql, mysql, redis-cli, mongosh) via subprocess. These are used to automate infrastructure management and database introspection. The scripts demonstrate good security awareness by reading from /dev/tty for user confirmation and passing sensitive headers via secure configurations.
[INDIRECT_PROMPT_INJECTION]: The database analysis functionality ingests database logs and SQL query text from the running environment for diagnostic purposes. Because this content originates from external application logs, it represents a surface for indirect prompt injection, where malicious instructions could be embedded in logs to attempt to influence the agent's analysis.
[OBFUSCATION]: Several Python scripts use Base64 encoding to wrap SQL query payloads before passing them to CLI tools. This is implemented as a functional measure to ensure query integrity and prevent shell command injection or parsing errors, rather than as a technique to hide malicious logic.

use-railway