docs: add host-side config and state file locations to README

[WuKongAI-CMU]

Summary

• add a README section that points users to the main host-side NemoClaw state files
• call out the optional service environment variables users are most likely to need
• steer reconfiguration back toward nemoclaw onboard instead of manual edits

Testing

• not run (docs-only change)

Fixes #494

Summary by CodeRabbit

• Documentation
  • Added "Host-Side State and Config" section explaining where NemoClaw stores operator configuration and state files on the host system.
  • Documented common environment variables for optional services and configuration.
  • Clarified recommended setup and reconfiguration procedures using the onboard command.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds a new documentation section to README.md explaining host-side state and configuration for NemoClaw. Documents three host file paths (~/.nemoclaw/credentials.json, ~/.nemoclaw/sandboxes.json, ~/.openclaw/openclaw.json), common environment variables, and recommends using nemoclaw onboard for setup rather than manual file editing.

Changes

Cohort / File(s)	Summary
Documentation README.md	Added "Host-Side State and Config" section documenting operator-facing state storage locations, host file paths with purposes, common environment variables, and configuration workflow recommendations.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Poem

🐰✨ A fluffy README update hops into view,
Config files documented, shiny and new!
No tangled secrets in sandbox walls—
Just home directory files that answer all calls!
Onboard with ease, the docs now say,
For operators, configuration's just a hop away! 🎉

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Linked Issues check	⚠️ Warning	The PR addresses issue #494 but incompletely: it documents file locations and environment variables, but omits the JSON example snippet and port configuration instructions from the original requirements.	Add JSON example showing how to configure channels.signal.httpPort and instructions for creating the file if it doesn't exist, as specified in issue #494.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and accurately describes the main change: adding documentation about host-side configuration and state file locations to the README.
Out of Scope Changes check	✅ Passed	All changes are directly related to documenting host-side configuration and state file locations as required by issue #494; no unrelated changes detected.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

🧹 Nitpick comments (1)

README.md (1)

205-218: Consider adding practical examples and workflow guidance.

The section successfully documents file locations, but issue #494 requested additional details that would help users actually work with these files:

1. JSON example: Show a snippet of ~/.openclaw/openclaw.json structure with a practical example (e.g., setting channels.signal.httpPort to 9090)
2. File creation: Note whether users need to create these files manually or if they're auto-generated
3. Environment variables: Explain where to set the listed env vars (shell profile, systemd service, etc.)
4. Post-change workflow: Clarify what happens after manual edits—does the user need to restart services, re-run nemoclaw onboard, or something else?

The current guidance "prefer nemoclaw onboard over editing these files by hand" is good advice, but users who do need manual edits would benefit from understanding the structure and consequences.

📝 Example enhancement

Consider adding after line 214:

Most of these files are created automatically during `nemoclaw onboard`. If you need to manually edit `~/.openclaw/openclaw.json`, it follows this structure:

\```json
{
  "channels": {
    "signal": {
      "httpPort": 9090
    }
  }
}
\```

After manual configuration changes, restart the relevant NemoClaw services with `nemoclaw stop` and `nemoclaw start`.

For environment variables, you might add after line 216:

Set these in your shell profile (`~/.bashrc` or `~/.zshrc`) or pass them when launching services:

\```bash
export TELEGRAM_BOT_TOKEN="your-token"
export ALLOWED_CHAT_IDS="123456,789012"
\```

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@README.md` around lines 205 - 218, Add practical examples and workflow
guidance for the host-side files: include a short JSON example showing the
structure of ~/.openclaw/openclaw.json (e.g., channels.signal.httpPort set to
9090), state whether ~/.nemoclaw/credentials.json and ~/.nemoclaw/sandboxes.json
and ~/.openclaw/openclaw.json are auto-created by the onboarding flow or must be
created manually (mention that most are created by running the comando 'nemoclaw
onboard'), document where to set TELEGRAM_BOT_TOKEN, ALLOWED_CHAT_IDS and
CHAT_UI_URL (shell profiles like ~/.bashrc, ~/.zshrc or systemd unit
environment) and add a clear post-edit workflow (e.g., after manual edits run
'nemoclaw stop' and 'nemoclaw start' or re-run 'nemoclaw onboard' as
appropriate) so users know the required restart/reload steps.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@README.md`:
- Around line 205-218: Add practical examples and workflow guidance for the
host-side files: include a short JSON example showing the structure of
~/.openclaw/openclaw.json (e.g., channels.signal.httpPort set to 9090), state
whether ~/.nemoclaw/credentials.json and ~/.nemoclaw/sandboxes.json and
~/.openclaw/openclaw.json are auto-created by the onboarding flow or must be
created manually (mention that most are created by running the comando 'nemoclaw
onboard'), document where to set TELEGRAM_BOT_TOKEN, ALLOWED_CHAT_IDS and
CHAT_UI_URL (shell profiles like ~/.bashrc, ~/.zshrc or systemd unit
environment) and add a clear post-edit workflow (e.g., after manual edits run
'nemoclaw stop' and 'nemoclaw start' or re-run 'nemoclaw onboard' as
appropriate) so users know the required restart/reload steps.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 7dc4e0c4-2118-4548-b132-d9c9622b4412

📥 Commits

Reviewing files that changed from the base of the PR and between 0ef5dd2 and fb4cb2a.

📒 Files selected for processing (1)

• README.md

[WuKongAI-CMU]

This is still approved on my side and I confirmed there are no follow-up changes pending from me. It is currently blocked by the base-branch merge policy for non-maintainers, so it looks ready for maintainer merge when convenient.

[WuKongAI-CMU]

Maintainer follow-up: this branch is still current, remains mergeable, and the host-side config/state path documentation is already in shape for merge. If there is any remaining policy or wording concern, I can patch it quickly.