runex
Pass
Audited by Gen Agent Trust Hub on Apr 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The
scripts/0.1.0/bundles.nuscript executes external system commandstarandcurlusing Nushell's native subprocess invocation.taris utilized to create compressed archives of workflow bundles, whilecurlhandles the transmission of these archives to the Runex API server. - [EXTERNAL_DOWNLOADS]: The management scripts (
runex.nu,debug.nu,bundles.nu) perform network communication with the Runex REST API usinghttp getandhttp postcommands. While the default target islocalhost:4001, the destination can be configured via environment variables to any remote host. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface through the retrieval and display of workflow step logs in
debug.nu. - Ingestion points:
scripts/0.1.0/debug.nuandscripts/0.1.0/runex.nuingest raw text data from the API response bodies (specifically theoutputanderrorfields of step runs). - Boundary markers: Absent. The scripts output the retrieved content directly to the console without delimiters or instructions to the agent to ignore embedded commands.
- Capability inventory: The skill allows subprocess execution (via
tarandcurl), network requests (viahttp), and file system interaction (ls,open,path). - Sanitization: None. The content retrieved from the API is displayed without escaping or filtering, which may lead an agent to interpret malicious instructions contained within retrieved log data.
Audit Metadata