indexes

Pass

Audited by Gen Agent Trust Hub on Jul 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill employs dynamic context injection (!date and !psql --version) to include system metadata when the skill is loaded. These commands are used for logging environment information and are considered benign usage of shell execution.
  • [PROMPT_INJECTION]: The skill performs a comprehensive scan of the codebase and incorporates user-supplied arguments, creating an attack surface for both direct and indirect prompt injection.
  • Ingestion points: The skill reads all files within the target codebase to identify SQL queries (SKILL.md).
  • Boundary markers: There are no specific delimiters or instructional warnings (e.g., "ignore embedded instructions") used to isolate the external codebase content.
  • Capability inventory: The agent has the capability to read files from the filesystem, write new documentation files, and interact with database command-line tools.
  • Sanitization: No evidence of content sanitization, filtering, or validation is present before the agent processes the code content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 26, 2026, 12:10 AM
Security Audit — agent-trust-hub — indexes