magic-linguistic-morph

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues were identified during the analysis of the skill's instructions, scripts, and data.
  • [EXTERNAL_DOWNLOADS]: The skill references several external linguistic resources (UniMorph, SIGMORPHON, HFST, Apertium). These are all reputable academic or open-source projects. No automated runtime downloads or execution of untrusted code were found.
  • [DATA_EXFILTRATION]: No patterns for data exfiltration were detected. The skill uses a local TSV snapshot for language code validation and does not perform network operations.
  • [COMMAND_EXECUTION]: The skill's scripts (paradigm_lookup.py, segmenter_recommend.py) are focused on data lookup and do not execute arbitrary shell commands or process untrusted input in a way that allows injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 07:00 AM
Security Audit — agent-trust-hub — magic-linguistic-morph