contracts-query-enhancer
Warn
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill utilizes the SN-Execute-Background-Script tool to run GlideScript on the ServiceNow server. In Step 6 and Example 3, it generates and executes JavaScript blocks to perform aggregations and logic checks. This constitutes dynamic code generation and execution from local templates.- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process natural language search queries from users and map them to database filters and scripts. This ingestion of untrusted data, combined with high-privilege execution tools, creates an attack surface where a crafted user query could potentially influence the resulting server-side script.
- Ingestion points: User-provided natural language queries (SKILL.md, Procedure Step 1).
- Capability inventory: SN-Execute-Background-Script, SN-Query-Table, SN-Create-Record.
- Boundary markers: The skill does not provide specific instructions to the agent on how to sanitize or delimit user input when interpolating it into script templates.
- Sanitization: No explicit sanitization or validation logic is provided for the natural language input before it is used to build queries or scripts.- [COMMAND_EXECUTION]: The skill includes instructions to use the Bash tool for performing REST API calls. While documented as examples for interacting with the ServiceNow API, the availability of this tool grants the agent the capability to execute shell commands.
Audit Metadata