postgresql
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides bash script snippets in gpt.md and SKILL.md intended for execution by the agent to perform database health checks and migrations using the psql utility. These scripts are standard administrative tools and are used within the context of the skill's primary purpose.
- [INDIRECT_PROMPT_INJECTION]: The skill enables the agent to ingest and analyze data from a PostgreSQL database, which could theoretically contain malicious strings if the database or processed queries are attacker-controlled. 1. Ingestion points: Result sets from system views like pg_stat_activity, pg_index, and EXPLAIN plan outputs (SKILL.md, gemini.md, gpt.md). 2. Boundary markers: Not explicitly implemented in the provided prompts. 3. Capability inventory: Shell execution (gpt.md) and SQL command execution through database tools. 4. Sanitization: No specific sanitization or filtering logic is provided for data retrieved from the database.
Audit Metadata