The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous examples of using the rpk command-line utility for linting, testing connections, and running pipelines. These are legitimate administrative tasks for the Redpanda Connect product.
[COMMAND_EXECUTION]: Includes curl commands directed at localhost:4195 to check service health endpoints and retrieve metrics. This is standard practice for monitoring local services.
[SAFE]: References standard configuration paths and environment variables such as /etc/redpanda/redpanda.license and REDPANDA_LICENSE. These are documented for the purpose of license management and secrets resolution, which is the intended functionality of the tool.
[PROMPT_INJECTION]: The skill involves processing external data via Bloblang mappings. Ingestion point: User-provided pipeline configuration files. Boundary markers: Absent. Capability inventory: Local command execution and network access. Sanitization: Handled by tool-specific validation logic. This represents a surface for indirect prompt injection, though no malicious patterns were identified in the skill content.

connect-debugging