The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the use of the b2c CLI tool to perform operations such as exporting, listing, and validating Page Designer content. This involves executing shell commands that interact with both the local filesystem and remote Salesforce instances.
[CREDENTIALS_UNSAFE]: The skill documentation describes how the b2c CLI tool accesses configuration files like dw.json and ~/.mobify to retrieve instance credentials. It also references the b2c setup inspect --unmask command, which is a standard feature of the developer tool used to reveal resolved secrets for troubleshooting. These are legitimate vendor practices.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from B2C Commerce library XML files and JSON metadefinitions, which serves as a potential attack surface.
Ingestion points: Data enters the agent context through files specified in commands like b2c content list --library-file and b2c content validate.
Boundary markers: No specific delimiters or instructions are provided to isolate this external content from agent instructions.
Capability inventory: The agent can execute CLI commands, read local files, and make network requests via the b2c tool.
Sanitization: The skill relies on the built-in parsing and validation logic of the underlying CLI tool.
[EXTERNAL_DOWNLOADS]: The skill recommends using npx @salesforce/b2c-cli if the tool is not installed globally. This command downloads code from a well-known service (NPM registry) and belongs to a trusted organization (Salesforce).

b2c-content