The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's preamble executes a series of shell commands and local binaries, such as rmbc-update-check and rmbc-analytics, to manage state and check for updates. It also uses shell utilities including grep, sed, mkdir, and touch to read and modify configuration files in the ~/.rmbc-skills/ directory.
[DATA_EXFILTRATION]: The skill implements a telemetry system that logs usage information, including the specific skill name and product identifiers, through a local binary.
[EXTERNAL_DOWNLOADS]: The skill provides a link to an external video on YouTube for user onboarding. YouTube is a well-known service.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface. Ingestion points: Untrusted data enters the agent context through the product_description and target_audience inputs. Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the ingested data. Capability inventory: The skill has access to shell execution, binary execution, and filesystem modification as defined in its preamble and post-execution steps. Sanitization: No sanitization or validation logic is applied to the user-provided inputs.

lead-writer