optimize-database
Pass
Audited by Gen Agent Trust Hub on Aug 16, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
- [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection by processing untrusted repository data for analysis. Ingestion points include database drivers, schemas, migrations, and log files as specified in SKILL.md. The skill lacks explicit boundary markers or instructions to ignore embedded prompts within audited files. Its capability inventory includes executing database queries and suggesting code modifications. No explicit sanitization of the input data is defined.
- [COMMAND_EXECUTION]: The skill executes database commands and query plans during the audit process. Evidence includes establishing performance baselines and tracing operations across various database engines. Safety mitigations are present, requiring explicit user confirmation before any destructive database operations.
- [CREDENTIALS_UNSAFE]: The skill audits sensitive connection configurations and credential storage locations. Evidence includes instructions in SKILL.md to search for connection configurations and audit credential rotation. This behavior is restricted to the primary auditing purpose and does not include patterns for external exfiltration.
Audit Metadata