shared-data
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill serves as a standard administrative resource for StarRocks database administrators, offering clear methodologies for identifying and resolving cloud-native architecture regressions.
- [COMMAND_EXECUTION]: The instructions utilize standard system tools such as
grep,curl,df, andsmartctlfor examining logs, metrics, and hardware health. These operations are restricted to the backend and frontend nodes of the cluster and are consistent with professional database maintenance. - [DATA_EXPOSURE_AND_EXFILTRATION]: The skill does not access sensitive user files or hardcode secrets. S3 configuration examples use descriptive placeholders like
<key>and<secret>, which is a safe practice. Network operations are directed at local cluster endpoints (e.g.,localhost:8040) for metric collection. - [PROMPT_INJECTION]: The skill processes diagnostic data from internal logs (
be.log,fe.log,be.WARNING). While these sources constitute an ingestion point for untrusted data, the skill does not interpolating this data into further automated prompts in a way that allows for instruction override. - [SAFE]: The skill incorporates a mandatory evidence chain for its diagnostic surfaces: 1. Ingestion points: StarRocks internal log files (SKILL.md); 2. Boundary markers: Absent (manual analysis); 3. Capability inventory: BE/FE configuration updates, cache directory deletion, and SQL administration (SKILL.md); 4. Sanitization: Absent (intended for manual administrator review).
Audit Metadata