meshcore-packet-capture

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill documents a remote installation pattern involving fetching and executing a shell script from a public GitHub repository.
  • Evidence: sudo bash -c "$(curl -fsSL https://raw.githubusercontent.com/agessaman/meshcore-packet-capture/main/install.sh)" found in references/deployment-and-troubleshooting.md.
  • [COMMAND_EXECUTION]: The skill provides instructions for executing various CLI tools and system management commands.
  • Evidence: Usage of meshcore-packet-capture, pipx install, systemctl, journalctl, and docker compose throughout SKILL.md and references/deployment-and-troubleshooting.md.
  • [PRIVILEGE_ESCALATION]: The instructions frequently require administrative privileges for installation, service management, and hardware access.
  • Evidence: Multiple uses of sudo for install.sh, nixos-rebuild, and systemctl. Requirement for privileged: true in Docker for Bluetooth (BLE) hardware access documented in references/deployment-and-troubleshooting.md.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to download resources from external domains during installation and runtime.
  • Evidence: Downloads the installer from raw.githubusercontent.com and references the upstream repository github.com/agessaman/meshcore-packet-capture.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes untrusted packet data from external radio sources.
  • Ingestion points: Captures data from MeshCore Companion radios via ble, serial, and tcp transports.
  • Boundary markers: The skill documents the use of structured output but does not mention specific sanitization for embedded instructions in packet payloads.
  • Capability inventory: The agent has capabilities to write to the filesystem via --output and publish data to external network endpoints via MQTT.
  • Sanitization: Not explicitly documented for the packet parsing logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 05:57 PM
Security Audit — agent-trust-hub — meshcore-packet-capture