model-supply-chain-security
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The
model_promotion_gate.pyscript accepts untrusted image strings as command-line arguments.\n - Ingestion points:
sys.argv[1]inmodel_promotion_gate.py.\n - Boundary markers: Absent.\n
- Capability inventory:
subprocess.runcalls tocosignandtrivyinmodel_promotion_gate.py.\n - Sanitization: Absent; image names are passed directly to the tool arguments.\n- [COMMAND_EXECUTION]: The skill uses Python's
subprocessmodule to execute external CLI tools.\n - Evidence:
subprocess.run()calls inmodel_promotion_gate.pyfor verification and scanning logic.\n- [EXTERNAL_DOWNLOADS]: The skill references and downloads tools and workflows from established sources.\n - Evidence: Uses the SLSA GitHub generator from the
slsa-frameworkorganization.\n - Evidence: Suggests installation of the
ficklingpackage for security analysis.
Audit Metadata