tessera

Warn

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install an R library using remotes::install_github("lassa-sentinel/GeoTessera"). This involves downloading code from a third-party GitHub repository that is not associated with a pre-verified trusted organization.
  • [REMOTE_CODE_EXECUTION]: The installation of the R library from a remote GitHub repository allows for the execution of code from an external source during the setup process.
  • [COMMAND_EXECUTION]: The skill makes extensive use of the geotessera CLI for downloading, visualizing, and serving data. Specifically, the geotessera serve ./web --open command launches a local web server, which increases the local network attack surface.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides the capability to ingest and process untrusted external data files, such as GeoJSON files via the --region-file flag.
  • Ingestion points: Data enters the system through the --region-file parameter in the CLI and list of coordinate tuples in the Python/R libraries (found in SKILL.md and references/geotessera-cli.md).
  • Boundary markers: No specific boundary markers or instructions to ignore embedded malicious content within the GeoJSON files are provided.
  • Capability inventory: The skill possesses the capability to write files to the local system (GeoTIFF, NPY, Zarr), execute shell commands via the CLI, and open a local web server.
  • Sanitization: There is no evidence of sanitization or validation of the internal metadata or structure of the GeoJSON files before they are processed by the underlying tools.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 19, 2026, 03:37 PM