Skills
skills/vladm3105/aidoc-flow-framework/doc-utest/Gen Agent Trust Hub

doc-utest

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local Python and Bash validation scripts (e.g., ai_dev_ssd_flow/10_TSPEC/scripts/validate_utest.py) to verify document structure and quality.
  • [PROMPT_INJECTION]: The skill processes local documentation files, which represents a surface for indirect prompt injection. 1. Ingestion points: files within docs/10_TSPEC/UTEST/. 2. Boundary markers: absent. 3. Capability inventory: local script execution via python and bash. 4. Sanitization: absent.
  • [SAFE]: The skill's behavior is consistent with its stated purpose of documentation management. No indicators of credential theft, data exfiltration, or obfuscation were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 04:26 AM
Security Audit — agent-trust-hub — doc-utest