total-recall

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime scripts call external LLM endpoints (e.g., the default LLM_BASE_URL https://openrouter.ai/api/v1 used by scripts/observer-agent.sh and the Anthropic endpoint https://api.anthropic.com/v1/messages used by scripts/backfill-importance.sh) and write model responses into observations.md which are then loaded into agent startup/system prompts, meaning remote responses are used at runtime to control agent instructions and are a required dependency.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.70). The skill's setup and quick-start explicitly run a setup script, create cron jobs, install packages (example shows "sudo apt install inotify-tools") and "install the reactive watcher service" (systemd), which modify the host filesystem and system services and may require elevated privileges.