The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides extensive documentation for VoltSP's ability to compile and execute Java code at runtime using the system compiler, as well as executing inline JavaScript and Python logic within processing pipelines (see references/pipeline-kinds/inline-java-in-yaml.md and references/plugins/processors/).
[INDIRECT_PROMPT_INJECTION]: The skill describes an architecture for ingesting and processing untrusted data from multiple external vectors, creating an attack surface where processed data could influence pipeline logic.
Ingestion points: Network sources (UDP/TCP), Kafka topics, MQTT subscriptions, and Elastic Beats (documented in references/plugins/sources/).
Boundary markers: Documentation encourages the use of secure configuration for secrets but lacks specific guidance on sanitizing untrusted inputs processed by scripting engines.
Capability inventory: Support for arbitrary code execution (Java/JS/Python), file system writes, and network exfiltration via dedicated sinks.
Sanitization: No explicit sanitization or input validation logic is present in the provided templates or documentation for handling remote data.
[COMMAND_EXECUTION]: The skill includes Python scripts (scripts/scaffold_voltsp_pipeline.py and scripts/check_voltsp_yaml_layout.py) that perform local file system operations, such as creating directory structures and writing template files based on user-provided arguments.
[EXTERNAL_DOWNLOADS]: The documentation references external resources for deployment and testing, including the official VoltSP Helm repository and Docker images (see references/kubernetes.md and references/testing.md).

voltsp