helmor-debug-operate

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill monitors and processes UI data such as accessibility trees and logs which may contain content from untrusted external sources, creating an indirect prompt injection attack surface.\n
  • Ingestion points: Data is ingested from the application webview using webview_dom_snapshot and read_logs (SKILL.md).\n
  • Boundary markers: The instructions lack explicit delimiters or warnings to treat UI-derived content as potentially untrusted data.\n
  • Capability inventory: The agent has access to highly capable tools including webview_execute_js, webview_interact, and webview_keyboard which can be used to alter the application state (SKILL.md).\n
  • Sanitization: Content retrieved from the application is not sanitized or validated before being used for subsequent agent reasoning.\n- [COMMAND_EXECUTION]: The skill employs custom JavaScript shims to interact directly with the application's internal messaging and element resolution systems.\n
  • A custom "Call App Commands" helper allows the agent to execute internal Tauri IPC commands by directly accessing window.__TAURI_INTERNALS__.ipc, bypassing limitations in the standard communication bridge.\n
  • The skill injects a "Selector Repair" script into the application's JavaScript environment to modify element resolution behavior at runtime.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 12:22 AM
Security Audit — agent-trust-hub — helmor-debug-operate