first-ever-not-first-in-window
Pass
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill suggests using
grepcommands to audit local Kotlin (.kt) files for specific SQL patterns such asMIN(timestamp),HAVING, andCOUNT(DISTINCT). These commands are standard for developer workflows and are scoped to the local environment for static analysis. - [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to process and analyze local source code to verify logical correctness.
- Ingestion points: Local source files (
.kt) identified and read viagrepcommands. - Boundary markers: Absent; the instructions assume the agent will parse the resulting file content directly.
- Capability inventory: Read-only shell access for file searching and inspection.
- Sanitization: None; the skill relies on the agent's built-in reasoning to evaluate the code fragments found.
Audit Metadata