The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the b2c CLI tool to perform various file system operations on remote B2C Commerce instances, including listing, downloading, uploading, and archiving files.
[CREDENTIALS_UNSAFE]: The documentation explicitly mentions the --unmask flag for the b2c setup inspect command. This flag reveals sensitive configuration secrets, such as client secrets and passwords, making them accessible to the agent context.
[EXTERNAL_DOWNLOADS]: The skill recommends using npx to download and execute the @salesforce/b2c-cli package. This involves fetching external code from the NPM registry at runtime, which is standard behavior for this official Salesforce tool.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection because it ingests and processes untrusted data from remote file systems.
Ingestion points: Remote file names and file contents are read into the agent's context via webdav ls and webdav get commands (documented in SKILL.md).
Boundary markers: No explicit delimiters or instructions to ignore embedded commands within remote files are specified.
Capability inventory: The skill provides full CRUD capabilities on the remote file system, including rm, put, and directory creation, as well as the ability to zip/unzip archives.
Sanitization: There is no mention of sanitization, validation, or escaping of the data retrieved from the remote WebDAV roots.

b2c-webdav