skills/nvidia/skills/cudaq-importing/Gen Agent Trust Hub

cudaq-importing

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process quantum circuit code from external frameworks (e.g., Qiskit). This creates a surface where instructions embedded within the source code data could potentially influence the agent's behavior during the porting process.
  • Ingestion points: The workflow in SKILL.md involves reading source circuit construction provided by the user.
  • Boundary markers: The skill provides a structured Port Validation Gate and explicit Core Rules to maintain algorithmic fidelity, which acts as a logical constraint.
  • Capability inventory: The skill uses Read, Glob, and Grep tools and generates Python code blocks.
  • Sanitization: The skill relies on technical gate mapping and comparison of count keys to ensure correctness rather than natural language interpretation of the source logic.
  • [DYNAMIC_EXECUTION]: The references/porting-reference.md file provides a code pattern (Source-Framework Import Blocker) that dynamically modifies the Python import system (sys.meta_path) to prevent the ported code from loading specified libraries. This is a standard validation technique for ensuring framework-free ports and does not pose a threat in this context.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 07:49 PM
Security Audit — agent-trust-hub — cudaq-importing