Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Register an agent email address on Rhobot Mail (name@rhobot.dev). Use when the user wants to set up agent email, register a Rhobot Mail account, claim an email handle, or onboard to rhobot.dev.
.claude/skills/mikeyobrien-rho-cloud-onboard/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 159% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 333% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 1431% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 75% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -17% | 0% |
Register an agent email address on Rhobot Mail and configure local credentials. The agent gets an inbox at name@rhobot.dev that can be polled via API.
tau for tau@rhobot.dev)https://rhobot.dev/emailhttps://rhobot.dev/skill.mdhttps://api.rhobot.dev (all API calls go here)curl and jq installedapi.rhobot.devAsk the user what email handle they want. If not provided, suggest one based on the agent's name or personality.
Constraints:
admin, postmaster, abuse, noreply, no-reply, support, help, info, security, webmaster, hostmaster, mailer-daemon, root, dmarc{handle}@rhobot.dev?"Before registering, check if credentials already exist.
bashcat ~/.config/rho-cloud/credentials.json 2>/dev/null
Constraints:
api_key, MUST warn the user: "Existing credentials found for {email}. Registering again will create a new account. Continue?"Call the registration endpoint.
bashcurl -s -X POST https://api.rhobot.dev/v1/register \ -H "Content-Type: application/json" \ -d '{"name": "{handle}", "display_name": "{display_name}"}' | jq .
Constraints:
ok field before proceeding| HTTP Status | Error | Action | |-------------|-------|--------| | 400 | Invalid name format | Show the validation rules. Ask user to pick a different handle. Return to Step 1. | | 409 | Handle already taken | Tell user {handle}@rhobot.dev is taken. Suggest alternatives: {handle}-agent, {handle}-bot, {handle}-ai, or ask user for another name. Return to Step 1. | | 500 | Server error | Report the error. Suggest retrying in a few minutes. | | Network error | Connection failed | Check internet. Suggest curl -s https://api.rhobot.dev/v1/health | jq . to verify the API is up. |
On successful registration, save the credentials locally.
bashmkdir -p ~/.config/rho-cloud cat > ~/.config/rho-cloud/credentials.json << 'EOF' { "api_key": "{api_key from response}", "agent_id": "{agent_id from response}", "email": "{email from response}" } EOF chmod 600 ~/.config/rho-cloud/credentials.json
Constraints:
Confirm the credentials work by hitting the health and status endpoints.
bash# Health check curl -s https://api.rhobot.dev/v1/health | jq . # Auth check curl -s -H "Authorization: Bearer {api_key}" \ https://api.rhobot.dev/v1/agents/status | jq .
Constraints:
{"ok": true}The registration response includes a claim_url for identity verification.
Constraints:
termux-open-urlSummarize what was set up.
Report template:
Registered: {handle}@rhobot.dev
Agent ID: {agent_id}
Credentials: ~/.config/rho-cloud/credentials.json
Claim URL: {claim_url}
Status: {claimed or pending}
Send a test email to {handle}@rhobot.dev to verify delivery.
Use the rho-cloud-email skill to check your inbox.| Problem | Solution | |---------|----------| | "already taken" on a handle you own | You may have registered previously. Check ~/.config/rho-cloud/credentials.json for existing credentials. | | Health check fails | API may be down. Try again in a few minutes. Check dig api.rhobot.dev for DNS resolution. | | Auth check returns 401 | API key may be invalid or corrupted. Re-register with a new handle. | | Claim URL returns 404 | Claim token may have expired. Re-register to get a fresh claim URL. |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→pass | 5,444 | 3,485 | -36% | 1 | 1 | 0% | 798 | 2,070 | +159% | 0 | 0 | — |
case-01 | fail→fail | 7,324 | 21,705 | +196% | 1 | 1 | 0% | 1,305 | 1,738 | +33% | 0 | 0 | — |
case-02 | fail→fail | 16,086 | 7,065 | -56% | 1 | 1 | 0% | 2,250 | 1,869 | -17% | 0 | 0 | — |
case-03 | fail→fail | 7,270 | 7,282 | +0% | 1 | 1 | 0% | 1,241 | 1,907 | +54% | 0 | 0 | — |
case-05 | fail→pass | 3,367 | 3,332 | -1% | 1 | 1 | 0% | 454 | 1,965 | +333% | 0 | 0 | — |
case-06 | pass→pass | 7,534 | 9,279 | +23% | 1 | 1 | 0% | 1,273 | 2,338 | +84% | 0 | 0 | — |
case-07 | fail→pass | 5,127 | 9,629 | +88% | 1 | 1 | 0% | 174 | 2,664 | +1431% | 0 | 0 | — |
case-08 | fail→pass | 8,503 | 7,121 | -16% | 1 | 1 | 0% | 1,576 | 2,751 | +75% | 0 | 0 | — |
case-09 | fail→pass | 13,530 | 2,956 | -78% | 1 | 1 | 0% | 2,351 | 1,949 | -17% | 0 | 0 | — |
case-10 | fail→pass | 9,115 | 2,049 | -78% | 1 | 1 | 0% | 1,620 | 1,745 | +8% | 0 | 0 | — |
case-11 | fail→pass | 6,903 | 1,753 | -75% | 1 | 1 | 0% | 1,163 | 1,663 | +43% | 0 | 0 | — |
case-12 | pass→pass | 11,214 | 3,453 | -69% | 1 | 1 | 0% | 1,935 | 1,992 | +3% | 0 | 0 | — |
case-13 | fail→pass | 8,317 | 1,703 | -80% | 1 | 1 | 0% | 1,433 | 1,672 | +17% | 0 | 0 | — |
case-14 | fail→pass | 10,591 | 3,634 | -66% | 1 | 1 | 0% | 1,593 | 1,944 | +22% | 0 | 0 | — |
case-15 | pass→pass | 3,825 | 3,389 | -11% | 1 | 1 | 0% | 683 | 2,146 | +214% | 0 | 0 | — |
case-16 | pass→fail | 6,336 | 1,960 | -69% | 1 | 1 | 0% | 1,047 | 1,759 | +68% | 0 | 0 | — |
case-17 | pass→pass | 4,918 | 1,837 | -63% | 1 | 1 | 0% | 829 | 1,635 | +97% | 0 | 0 | — |
case-18 | pass→pass | 4,606 | 3,892 | -16% | 1 | 1 | 0% | 737 | 2,107 | +186% | 0 | 0 | — |
case-19 | fail→pass | 6,631 | 3,402 | -49% | 1 | 1 | 0% | 1,103 | 2,000 | +81% | 0 | 0 | — |
case-20 | fail→pass | 3,198 | 13,250 | +314% | 1 | 1 | 0% | 521 | 3,159 | +506% | 0 | 0 | — |
case-21 | fail→fail | 2,806 | 27,786 | +890% | 1 | 1 | 0% | 469 | 6,022 | +1184% | 0 | 0 | — |
case-22 | fail→pass | 9,794 | 5,268 | -46% | 1 | 1 | 0% | 1,771 | 2,366 | +34% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +50 percentage points is the difference between those two pass rates over the 18 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.