huawei-cloud-openviking-embedding-switch

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes system commands such as kill to terminate existing server processes and rm to clear local vector database indexes. These actions are performed to ensure a clean state when changing embedding dimensions, which is consistent with the skill's stated purpose.
  • [EXTERNAL_DOWNLOADS]: The skill uses curl to interact with local services hosted on 127.0.0.1 (ports 8090 and 1933). These interactions are used to query environment state and send execution requests to a local sandbox manager.
  • [REMOTE_CODE_EXECUTION]: Automated scanners flagged patterns involving curl piped to python3. However, technical analysis shows that these commands use the python3 -c flag with hardcoded JSON parsing logic to process the output of local APIs. The data being piped is JSON configuration or state data, not executable code from a remote source, making this a safe method for data extraction.
  • [DATA_EXPOSURE]: The skill reads and modifies local configuration files (ov.conf) and metadata (collection_meta.json) within a specific sandbox directory. These operations are restricted to the local environment and do not involve the transmission of sensitive data to external domains.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 12:19 PM
Security Audit — agent-trust-hub — huawei-cloud-openviking-embedding-switch