momentic-test
Pass
Audited by Gen Agent Trust Hub on Nov 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The test execution environment preloads the
child_processmodule, allowing the agent to execute arbitrary shell commands during test runs. - Evidence: Found in
references/sessions.mdwhich listschild_processas a preloaded library for Node and mobile JavaScript. - [DYNAMIC_EXECUTION]: The skill is designed to execute JavaScript code embedded within
*.test.yamland*.module.yamlfiles to handle setup, data generation, and complex assertions. - Evidence:
references/products/web/authoring.mdandreferences/products/mobile/authoring.mddescribe using JavaScript for data generation, APIs, and databases, noting that bare YAML string inputs are often evaluated as JavaScript. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits a vulnerability surface by ingesting and executing instructions from YAML files, which could lead to malicious code execution if the files originate from untrusted sources.
- Ingestion points:
*.test.yamland*.module.yamlfiles (mentioned inSKILL.md). - Boundary markers: Absent; there are no explicit delimiters or warnings to ignore embedded instructions within the ingested YAML files.
- Capability inventory: Shell access via
child_process, network requests viaaxios, and database interactions viapg(listed inreferences/sessions.md). - Sanitization: Absent; the skill executes JavaScript expressions found in the YAML files directly through the provided testing CLI tools.
Audit Metadata