fix-slow-queries
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill documents best practices for managing cloud database credentials, specifically warning against insecure storage or logging of connection strings and secrets. It advocates for read-only diagnostic connections to minimize operational risk.
- [COMMAND_EXECUTION]: The workflow involves running local benchmarking tools and Git commands to establish performance baselines. These operations are performed using the vendor's established tooling (e.g., @paperclipai/db) and are confined to the local development environment.
- [DYNAMIC_EXECUTION]: The skill uses dynamic module loading to compare code across different Git revisions for benchmarking purposes. This behavior is restricted to the local filesystem and intended strictly for performance regression testing.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data such as database query logs and statistics. It addresses potential risks by requiring data sanitization and focusing on performance metrics rather than automated execution of log-derived instructions.
Audit Metadata