external-gitcode-ascend-ascendc-operator-code-gen

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes several shell commands to build and verify the operator code, including bash build.sh, pip install for the generated wheel file, and running Python/Pytest scripts for testing.
  • [PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection. It ingests untrusted data from ascend-kernel/csrc/ops/<op_name>/design.md and uses it to generate executable C++ and Python code. An attacker could potentially embed malicious instructions in the design document that result in the generation of harmful code or exfiltration of sensitive information during the compilation or testing phase.
  • Ingestion points: ascend-kernel/csrc/ops/<op_name>/design.md (SKILL.md)
  • Boundary markers: None identified in the prompt templates or instructions for processing the design document.
  • Capability inventory: bash build.sh, pip install, and python tests/test_<op_name>.py in SKILL.md.
  • Sanitization: No explicit sanitization or validation of the input logic from the design document is described.
  • [COMMAND_EXECUTION]: The skill instructs the agent to modify file permissions using chmod +x build.sh to ensure the build script is executable.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 03:03 AM
Security Audit — agent-trust-hub — external-gitcode-ascend-ascendc-operator-code-gen