plan-rls-audit

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a defensive tool focused on security auditing. It explicitly forbids the execution of SQL commands or database modifications, emphasizing a 'plan-only' approach where the final output is a markdown report.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of database schemas, migration files, and policy definitions. While these sources could theoretically contain adversarial text, the skill's instructions are focused on structural auditing and reporting, which limits the potential impact of such injections. No specific boundary markers are mentioned for the ingested schema data, but the context is purely analytical.
  • [DATA_EXPOSURE]: The instructions direct the agent to search for leaked credentials (e.g., service_role keys) in the user's frontend code. This is a standard security auditing practice aimed at identifying and remediating existing vulnerabilities, rather than an attempt to exfiltrate new data.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 11:26 PM