Letting an AI answer WhatsApp is convenient until it says something off-brand or unsafe to a customer. This assistant wraps every reply in guardrails that block off-limits topics, rate-limit abuse, and catch scam patterns.
sonnet5 daysWhatsApp Business APIClaudeSupabaseTwilio
Claude
74ROI
82Scale
$2.8k93Saved
ROI for
README.md
Why this subagent
Letting an AI answer WhatsApp is convenient until it says something off-brand or unsafe to a customer. This assistant wraps every reply in guardrails that block off-limits topics, rate-limit abuse, and catch scam patterns.
It answers within a defined scope in your tone, and hands off to a human the moment a request crosses the line. You get fast, on-message replies on the channel customers prefer without handing them an unguarded bot.
How it runs
Used at step 01 to kick off the pipeline.
Write
Used at step 01 to kick off the pipeline.
WebFetch
Used at step 01 to kick off the pipeline.
WebSearch
Used at step 01 to kick off the pipeline.
Check guardrails for blocked topics, rate limits, and known scam patterns.
pending
Generate a reply within the allowed scope and the brand's tone.
pending
Send the response and hand off to a human when the request crosses a guardrail.
pending
Sample output
json
// Sample output
// (generated when the pipeline finishes)
Reply to the WhatsApp message within the allowed topics; if it breaks a guardrail, reply HANDOFF only.
Unlock the rest
The full agent definition, install snippet, and starter task are gated for community members.
Members get the full `.md` agent file, the npm / pnpm install one-liners, a starter prompt that we've tuned against real runs, and the open-source repo when this automation ships there. One email, magic link, done.