amw-memory-operator
Pass
Audited by Gen Agent Trust Hub on Feb 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions and templates for executing local shell commands via
npm run amwto control the automation workbench and manage session profiles. - [REMOTE_CODE_EXECUTION]: The
eval_jsaction enables the execution of arbitrary JavaScript code within the context of the automated browser session, which is a powerful runtime execution capability. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface (Category 8):
- Ingestion points: The agent ingests untrusted data from external websites via actions like
snapshot,eval_js, andget_url. - Boundary markers: No explicit delimiters or boundary markers are defined to isolate external web content from the agent's internal instruction context.
- Capability inventory: The agent can execute JavaScript in the browser environment, write artifacts to the local filesystem, and trigger local CLI operations.
- Sanitization: No evidence of sanitization, filtering, or validation of the content retrieved from websites is present before it is processed by the agent.
Audit Metadata