skills/joelhooks/joelclaw/executor/Gen Agent Trust Hub

executor

Pass

Audited by Gen Agent Trust Hub on Jul 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: Executes administrative shell commands and requests sudo privileges to install and manage system-level LaunchDaemons and verification scripts.
  • [EXTERNAL_DOWNLOADS]: Downloads source code from a well-known public GitHub repository (RhysSullivan/executor) for the purpose of validating service logic and configuration.
  • [DATA_EXFILTRATION]: Configures network services to bind to all interfaces (0.0.0.0), facilitating legitimate remote access via LAN and Tailscale as part of the intended service contract.
  • [PROMPT_INJECTION]: Ingests data from service logs (Ingestion: /Users/Shared/joelclaw/logs/executor/launchd.out.log) using tail. Boundary markers: Not present. Capabilities: Administrative access (sudo) and system service management. Sanitization: None explicitly defined, although the skill includes specific rules against exposing sensitive authentication files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 5, 2026, 05:04 PM
Security Audit — agent-trust-hub — executor