elasticsearch-ingest
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on the
elasticCLI to interact with Elasticsearch clusters. It maps HTTP-shorthand operations to specific CLI commands such aselastic es bulk,elastic es indices create, andelastic es indices deletefor managing indices and ingesting data. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and parses untrusted local data (CSV, JSON, NDJSON) to convert it into a bulk-loadable format. It does not provide specific instructions to use boundary markers or to sanitize the content of these files against embedded malicious instructions.
- Ingestion points: Reads and samples local CSV and JSON files during the classification and conversion steps (SKILL.md, Step 2).
- Boundary markers: None specified in the instructions for delimiting untrusted data during processing.
- Capability inventory: File system access for reading source files and writing temporary NDJSON files, combined with network operations through the
elasticCLI (SKILL.md, Operations table). - Sanitization: No explicit sanitization, escaping, or validation logic is defined for the data values extracted from files before they are included in the bulk ingest payload.
Audit Metadata