How to use solid agents Mila, Sloane, Elon:
- Mila: "Mila, creative direction and visuals..." (design, CX, assets)
- Sloane: "Sloane, draft content and SEO..." (writing, satire, optimization)
- Elon: "Elon, bold first-principles execution and ship..." (planning, deploys)
- Combine with Hermes for execution muscle: "Mila creative, Sloane content, Elon bold plan, Hermes deploy"
- Be concrete: request specific outputs like "generate thumbnails", "write blog post", "create deploy plan".
- Follow OpenClaw: per-agent IDENTITY.md / AGENTS.md / config-agents, read-first for safety, bindings for channels. Hosted: Pure Cloudflare (Pages/Worker). No Vercel. Access control via Cloudflare Zero Trust recommended for privacy.
Communication Webapp (SMS / GChat / Slack / WhatsApp via Hermes)
This mini webapp fills a Hermes task for real execution (local Hermes required for actual sends). Cool tasks: client reminders, team digests, alerts.
Active Agent Team (click to toggle • click [solo] for single-agent command — defaults to one)
Defaults to single agent (Samantha) — perfect for "just give one agent a command". Click any agent to toggle (builds a team). Click [solo] next to an agent for instant single-agent mode. Mila/Sloane/Elon = solid core specialists. Hermes = execution powerhouse.
Invoke directly: "Mila, create the visuals...", "Sloane, draft the content...", "Elon, bold plan the rollout...". Pair with Hermes for full execution (deploys, terminal, automation).
HERMES SUPERPOWERS (v0.18+ Agent CLI — the execution engine):
Local: ~/.local/bin/hermes or `hermes skills list`, `hermes mcp`, `hermes computer-use`, `hermes claw`, `hermes serve --tui` etc. Invoke with "Hermes, use skills + mcp + computer-use + claw...". Ties to OpenClaw via claw subcommand. Isolated but powerful when authorized.
Per OpenClaw config-agents + Hermes runtime. Identity from IDENTITY.md. Mila, Sloane, Elon = solid reliable specialists. Hermes executes for real. Boardroom on Cloudflare. Private via Cloudflare Access.