benzinga-automation
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (HIGH): The skill instructs users to add an external MCP endpoint (
https://rube.app/mcp). Because this domain is not within the trusted scope (TRUST-SCOPE-RULE), this is classified as a high-severity finding regarding untrusted external code sources.\n- [REMOTE_CODE_EXECUTION] (HIGH): The use ofRUBE_REMOTE_WORKBENCHimplies the ability to execute logic on a remote server. This capability, combined with an untrusted source, poses a significant risk of arbitrary execution.\n- [PROMPT_INJECTION] (HIGH): The skill exhibits a high-severity indirect prompt injection surface (Category 8).\n - Ingestion points: External financial data and toolkit schemas ingested via
RUBE_SEARCH_TOOLSand Benzinga API calls inSKILL.md.\n - Boundary markers: Absent. There are no delimiters or instructions to ignore embedded commands in the data retrieved from Benzinga.\n
- Capability inventory: High. The agent can use
RUBE_MULTI_EXECUTE_TOOLandRUBE_REMOTE_WORKBENCHto perform side-effect-heavy actions.\n - Sanitization: Absent. No evidence of input validation or content filtering is present in the skill definition.
Recommendations
- AI detected serious security threats
Audit Metadata