host-triagelisted

# Host Triage Use this skill before making changes on a target host or when cached host facts may be stale. ## Triage Checklist - Load scoped Skill runtime memory first, then verify stale or high-impact facts with read-only probes before relying on them. - If the target OS, shell, gateway, join URL, workspace path, service manager, installed agent framework, or permission level is unclear, inspect the environment first and ask a concise follow-up before choosing commands. - OS, version, architecture. - User identity and admin/elevation status. - Shell and PowerShell availability. - Git, SSH, package manager, Codex, Claude, acpx. - Network reachability to the configured Remote Dev Skillkit gateway or join URL. - Proxy/DNS state, NAT/firewall/CGNAT constraints, SSH configuration, installed tunnel or mesh tools, and available connection modes. - Workspace boundaries. - Existing security tools or enterprise restrictions. - Installed service managers or launch surfaces: LaunchAgent/launchd, systemd user units, Windows Service Control Manager, scheduled tasks, or foreground shell only. - Existing `rdev` binary, Skillkit files, MCP configuration, and relevant environment variables. - If `rdev` is missing, record it as recoverable before escalating: look for an active checkout, current executable path, Go toolchain, and Git clone ability; use `rdev bootstrap agent-plan --repo-root .` when available or `go run ./cmd/rdev bootstrap agent-plan --repo-root .` from