Problem

The OpenClaw migration currently lives as an optional skill (optional-skills/migration/openclaw-migration/). As a skill, it requires the agent to already be running — which means the user needs working API keys and platform configs set up first. This is a chicken-and-egg problem since the migration is supposed to bring those keys and configs over.

Proposal

Add a hermes migrate --from openclaw CLI subcommand that:

1. Runs before agent startup (no API keys needed)
2. Auto-detects ~/.openclaw/ and offers to migrate
3. Reuses the existing openclaw_to_hermes.py script directly
4. Presents an interactive flow (preset selection, conflict modes, secret migration opt-in)
5. Could optionally be offered during hermes setup if ~/.openclaw/ is detected

Why this matters

With the expanded migration (19 categories now), the script can bring over essentially everything a user needs to get Hermes working: API keys, platform tokens, model selection, TTS config, allowed users, skills, memories, and persona. Running this before first launch means the user gets a fully configured Hermes on day one.

Implementation notes

• Add migrate subcommand to hermes_cli/main.py argparse
• --from flag accepts openclaw (extensible to other agents later)
• Interactive prompts via stdlib input() (no agent/prompt_toolkit needed)
• The migration script at optional-skills/migration/openclaw-migration/scripts/openclaw_to_hermes.py already handles everything — just need a CLI wrapper
• Consider auto-detection in hermes setup: if ~/.openclaw/ exists, ask "Would you like to import your OpenClaw settings?"
• Could also detect other agent configs (~/.claude/, ~/.codex/) for future migration paths

Current state

The migration script supports: SOUL.md, MEMORY.md, USER.md, AGENTS.md, workspace + shared skills, command allowlists, Telegram/Discord/Slack/WhatsApp/Signal settings, model config, TTS config, provider API keys, TTS asset files, daily memory notes, and archiving unmapped docs. 19 categories total with 24 tests.