The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided log samples (text, JSON, ndjson) to generate structured pipeline configurations.
Ingestion points: Phase 2 specifies that the agent should ask the user for sample input data.
Boundary markers: The skill does not explicitly define boundary markers or delimiters to isolate user-provided data from the agent's instructions.
Capability inventory: The skill utilizes the dryrun_pipeline tool from the greptimedb-mcp-server to test generated configurations.
Sanitization: No explicit sanitization or filtering of the user-provided sample data is mentioned during the transformation process.
[DYNAMIC_EXECUTION]: The skill instructs the agent to generate YAML configurations that include VRL (Vector Remap Language) scripts for complex data processing.
Script generation: The skill provides templates for creating vrl processor blocks within the pipeline YAML.
Execution method: The generated scripts are intended for execution within the GreptimeDB pipeline environment or via the dryrun_pipeline MCP tool.
[SAFE]: References to external documentation target docs.greptime.com and vector.dev, which are official and well-known service domains for the respective technologies mentioned in the skill. The primary function of the skill is to serve as a configuration assistant, and all mentioned tools (e.g., dissect, regex, date processors) are standard components of the GreptimeDB logging system.

greptimedb-pipeline