message-broker
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill describes an AHKBridge module that includes a run_script method for executing AutoHotkey scripts. This allows for local command execution as part of the intended inter-process communication logic.
- [PROMPT_INJECTION]: The broker architecture facilitates the exchange of arbitrary JSON payloads between processes. This creates an indirect prompt injection surface where malicious data ingested from external sources could be published to topics, potentially triggering unintended actions in subscribing components. Ingestion points: TCP socket on port 15234 (SKILL.md). Boundary markers: JSON-RPC structure (present). Capability inventory: run_script functionality in AHKBridge and message publishing (SKILL.md). Sanitization: None demonstrated in the protocol or implementation examples.
Audit Metadata