mac-agent-runs

Pass

Audited by Gen Agent Trust Hub on Oct 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides the kilobat awake -- <command> pattern, which allows for the execution of arbitrary shell commands. While its intended use is to wrap legitimate tasks like builds or downloads to prevent system sleep, it represents a generic subprocess execution capability.
  • [EXTERNAL_DOWNLOADS]: The instructions suggest installing the kilobat CLI via brew install seifip/kilobat/kilobat. This involves downloading and installing software from an external repository (Homebrew tap).
  • [INDIRECT_PROMPT_INJECTION]: The skill defines tools that ingest and process data from external network sources and local session logs, creating a potential surface for indirect injection.
  • Ingestion points: Data retrieved by list_sessions (including network destinations and events) and diagnose_connection (which processes network responses, DNS, and TLS data from user-provided target hosts).
  • Boundary markers: The skill does not define explicit delimiters or instructions to treat data from these tools as untrusted content.
  • Capability inventory: The skill has the ability to execute shell commands via the kilobat wrapper and modify system power management behavior.
  • Sanitization: There is no evidence of sanitization or validation of the data returned by connection diagnostics or session history before it is presented to the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 29, 2026, 01:24 PM