fluent-sdk
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data retrieved from ServiceNow instance tables which could influence agent behavior.
- Ingestion points: Reads data from ServiceNow tables such as
sys_update_xmlandsys_userusing theSN-Query-Tabletool (described in SKILL.md, Step 6). - Boundary markers: Absent; there are no specific instructions to delimit or ignore instructions embedded within the record data.
- Capability inventory: The skill utilizes shell execution (Bash), file writing, and server-side script execution via
SN-Execute-Background-Script(SKILL.md). - Sanitization: Absent; no explicit validation or escaping of the ingested ServiceNow data is described before use.
- [EXTERNAL_DOWNLOADS]: Installs the
@servicenow/sdkpackage from the official NPM registry and references documentation from the official ServiceNow GitHub repository. These represent standard dependencies from a well-known service provider. - [COMMAND_EXECUTION]: Executes system commands via Bash to verify the development environment (
node --version) and manage the application lifecycle using CLI tools (now-sdk build,now-sdk deploy). - [DYNAMIC_EXECUTION]: Assembles and executes ServiceNow background scripts and fix scripts via MCP tools (
SN-Execute-Background-Script,SN-Create-Fix-Script) and uses theNow.include()helper for dynamic script inclusion within Fluent project definitions.
Audit Metadata