bailian-docs-llm-wiki
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill contains no executable scripts or logic. It is composed entirely of documentation in Markdown format and structured data in JSON/JSONL format for knowledge retrieval purposes.
- [SAFE]: Hardcoded credential patterns were checked. All sample code provided for user reference uses secure placeholders like 'sk-xxx', 'st-****', or retrieves keys from environment variables (e.g.,
os.getenv("DASHSCOPE_API_KEY")), adhering to standard security best practices. - [SAFE]: All external URL references and network commands (curl/wget examples) point to legitimate Alibaba Cloud infrastructure and established service domains (e.g.,
aliyuncs.com,aliyun.com), which are recognized as well-known and trusted technology services. - [SAFE]: Instructional content includes prompt engineering templates for specific use cases (e.g., robot control). These are clearly identified as educational examples for users and do not contain instructions aimed at subverting the agent's own safety protocols or system prompts.
- [SAFE]: The skill uses
jqandgrepcommands in its instructions for searching local data files. These operations are performed on the skill's own data and do not involve untrusted input or risky execution paths.
Audit Metadata