ExUnit Test Framework
Fail
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION] (HIGH): The
generate-test.exsscript is vulnerable to Elixir code injection. It uses string interpolation to embed user-provided arguments (module,description) directly into a generated.exsfile. Because Elixir scripts are executed during testing via therun-test.exsscript, an attacker can escape the module context to run arbitrary system commands. - [PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection when used by other agents (like the mentioned
deep-debugger) to process untrusted bug reports. • Ingestion points: The--moduleand--descriptionflags ingenerate-test.exs. • Boundary markers: None; input is placed directly into code templates. • Capability inventory: File system write access (File.write!) and command execution (System.cmd). • Sanitization: None; the script performs no escaping or validation of inputs. - [COMMAND_EXECUTION] (MEDIUM): The skill allows arbitrary file writes via the
--outputparameter ingenerate-test.exs. An attacker could specify sensitive paths (e.g.,~/.bashrcor system configuration files) to overwrite them with generated code, leading to system instability or denial of service.
Recommendations
- AI detected serious security threats
Audit Metadata