The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs users to install the CLI by downloading a script from cli.tilebox.com and piping it directly to a shell (curl | sh). This is a common but sensitive pattern for software installation. The domain is owned by the skill's vendor.
[COMMAND_EXECUTION]: The skill utilizes the tilebox command-line tool and jq for data processing. This includes the tilebox parallel command, which is designed to execute other shell commands as arguments.
[DATA_EXFILTRATION]: The skill manages authentication using the TILEBOX_API_KEY environment variable. It includes explicit instructions to avoid printing or logging these credentials, which aligns with security best practices for credential management.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface (Category 8) because it ingests untrusted data from external sources and incorporates it into the agent's context.
Ingestion points: Data enters the context via tilebox docs search, which retrieves excerpts from docs.tilebox.com, and tilebox agent-context, which retrieves command metadata and schemas.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content embedded in the documentation search results or schemas.
Capability inventory: The agent has the capability to execute shell commands via the tilebox CLI and perform parallel task execution.
Sanitization: No sanitization or validation of the retrieved documentation content is mentioned before it is processed by the agent.

using-tilebox-cli