ysql

Pass

Audited by Gen Agent Trust Hub on Oct 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and process user-supplied SQL queries, schema definitions, and application code as defined in the SKILL.md frontmatter. This creates an ingestion surface for potentially malicious instructions embedded in user data. The documentation includes markdown structure and code blocks as boundaries and recommends the use of prepared statements for sanitization. The skill's capabilities include executing SQL through drivers, managing files via the yb-voyager tool, and performing network operations for diagnostics.
  • [EXTERNAL_DOWNLOADS]: The instructions include commands to download and install various YugabyteDB-specific drivers and tools from official registries such as PyPI, NPM, and RubyGems. Examples include packages like psycopg-yugabytedb, @yugabytedb/pg, yugabytedb-ysql, and yb-postgres across the reference files.
  • [COMMAND_EXECUTION]: The skill provides several shell commands for tool installation and environment verification. A diagnostic command in references/voyager.md uses curl to fetch server configuration from a local tablet server: curl -s http://<tserver>:9000/varz?raw | grep ysql_sequence_cache_minval.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 19, 2026, 06:21 PM