The Agent Skills Directory

[SAFE]: Comprehensive analysis of the skill's instructions and source code identified no malicious behaviors or security risks. The skill operates as a standard scientific data processing utility.
[EXTERNAL_DOWNLOADS]: The skill documentation indicates that the dara-xrd library may automatically download the BGMN software binary if a local version is not provided. This is a legitimate and documented setup process for the DARA framework.
[COMMAND_EXECUTION]: The refinement process involves executing scientific logic and interacting with refinement binaries (BGMN) via the DARA library. This activity is the primary intended function of the skill for materials characterization.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface through the ingestion of user-provided scientific data files.
Ingestion points: XRD patterns and CIF structure files are read in scripts/refine.py and scripts/convert_xrd_to_xy.py.
Boundary markers: No explicit boundary markers or instruction-ignoring warnings are used during file ingestion.
Capability inventory: The skill has capabilities to write results to the file system and execute refinement algorithms.
Sanitization: The data is parsed into numeric arrays and structural models; however, no natural language sanitization is applied. The specialized technical format of these files significantly limits the feasibility of prompt injection.

mat-xrd-refinement