query-budget-enforcer
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill consists of valid instructional content for database query optimization. No patterns of overriding agent safety guidelines, jailbreaking, or system prompt extraction were detected.
- [DATA_EXFILTRATION]: While the skill mentions connecting to database APIs for metrics, no hardcoded credentials, sensitive file path access, or unauthorized network operations were identified. The output is directed to local documentation paths.
- [COMMAND_EXECUTION]: There are no shell commands, subprocess calls, or system-level modifications present in the skill files.
- [EXTERNAL_DOWNLOADS]: No remote scripts or packages are downloaded or executed. The skill relies on its internal markdown logic and the agent's pre-existing database capabilities.
- [REMOTE_CODE_EXECUTION]: No remote execution patterns or dynamic code evaluation methods (like eval or exec) were found.
- [INDIRECT_PROMPT_INJECTION]: The skill processes SQL queries which are external data. While this represents a potential ingestion surface for malicious instructions hidden in SQL comments, the skill is tasked with textual analysis and reporting, which limits the impact of such injections. Standard LLM guardrails apply to the analysis phase.
Audit Metadata