external-gitcode-ascend-ascendc-operator-doc-gen

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local source files, including csrc/register.cpp, csrc/ops.h, design.md, and others. This represents an indirect prompt injection surface where maliciously crafted content in these files (such as in documentation comments or design notes) could attempt to influence the agent's output. However, the skill's logic is constrained to specific extraction and formatting patterns for documentation generation.
  • Ingestion points: Multiple source and design files in the csrc/ and tests/ directories.
  • Boundary markers: None explicitly defined in the prompt to separate data from instructions.
  • Capability inventory: Writing to local files (ascend-kernel/csrc/ops/<op_name>/README.md) and displaying content in the chat interface.
  • Sanitization: The skill relies on natural language extraction rules without explicit sanitization mechanisms.
  • [EXTERNAL_DOWNLOADS]: The skill is synchronized from https://gitcode.com/Ascend/agent-skills, which is a repository associated with the Ascend AI platform. This reference is documented neutrally as a source of the skill's instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 03:03 AM
Security Audit — agent-trust-hub — external-gitcode-ascend-ascendc-operator-doc-gen