The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill performs a shallow clone of the GreptimeDB source code from the official vendor repository (https://github.com/GreptimeTeam/greptimedb). This is used to map user descriptions to specific log strings and source modules for more accurate discovery. All clones are performed into temporary directories and intended for read-only analysis.
[COMMAND_EXECUTION]: The skill executes provided shell and Python scripts for data retrieval and validation:
scripts/http_sql_query.sh and scripts/http_arrow_export.sh: Use curl to interface with the GreptimeDB HTTP API. These scripts include robust internal validation to reject SQL comments, multiple statements, or non-read-only queries.
scripts/validate_arrow.py: A Python utility used to verify the integrity and row counts of exported Arrow IPC files.
Authentication headers are managed via temporary configuration files to prevent credential exposure in process listings.
[DATA_EXFILTRATION]: The primary purpose of the skill is to export database telemetry to the user's local environment. It manages this through several secure modes (HTTP export, COPY TO, or kubectl cp), all of which require explicit user confirmation for window selection and large-scale data transfers.

self-monitoring-export