← Back to Skills Marketplace
chris-openclaw

Google Business Review Responder

by chris-openclaw · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
98
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install google-review-responder
Description
Monitors new Google Business reviews, drafts tailored responses sent via Telegram for operator approval before posting publicly.
README (SKILL.md)

Skill: Google Business Review Responder

Description

Automatically monitors Google Business Profile reviews for clients, drafts professional responses, and sends them via Telegram for approval before posting.

Trigger

This skill activates during heartbeat checks and when the operator replies to a review approval message.


Review Check Flow (Heartbeat)

  1. Run the review check script for each configured client:
    python3 ~/review-responder/gbp_reviews.py check --client \x3Cclient_id>
    
  2. For each new unanswered review found, draft a response following the Response Guidelines below.
  3. Send the draft to the operator via Telegram in this format:
📝 New Review for [Business Name]

⭐ [star_rating] from [reviewer_name]
💬 "[review comment]"

My draft reply:
"[your drafted response]"

Reply OK to post, or send your edits.
(Review ID: [review_id] | Client: [client_id])
  1. Do NOT post the reply automatically. Wait for operator approval.

Approval Flow (Chat)

When the operator replies to a review draft:

  • "OK" or "post it" or "send it" or "approved": Post the draft as-is using:

    python3 ~/review-responder/gbp_reviews.py reply --client \x3Cclient_id> --review \x3Creview_id> --reply "the approved response"
    

    Confirm once posted: "Done -- reply posted for [reviewer_name]'s review."

  • Edited text: If the operator sends replacement text (anything that isn't a simple approval), use their text as the reply instead. Confirm before posting: "Got it -- posting your version now."

  • "Skip" or "ignore": Do not reply to that review. Remove it from pending.


Response Guidelines

Tone Principles

  • Warm, professional, and human -- not corporate or robotic
  • Specific to what the reviewer said (never generic "thanks for your review!")
  • Concise: 2-4 sentences max
  • Match the energy of the review without being over the top

By Star Rating

5 Stars:

  • Thank them warmly and reference something specific they mentioned
  • Reinforce what they loved ("We're glad [specific thing] made a difference")
  • End with a light invitation to return or share with others
  • Keep it brief -- don't overdo it on a great review

4 Stars:

  • Thank them and acknowledge specific positives
  • If they mentioned something that could improve, acknowledge it gracefully without being defensive
  • Show you're listening: "We appreciate the feedback on [topic] and are always looking to improve"

3 Stars:

  • Thank them for taking the time
  • Acknowledge both the positives and the concern
  • Show genuine interest in making it right: "We'd love the chance to do better next time"
  • Optionally invite them to reach out directly

1-2 Stars:

  • Lead with empathy, not defensiveness: "We're sorry to hear this wasn't the experience you deserved"
  • Acknowledge the specific issue without making excuses
  • Offer a path forward: invite them to contact the business directly
  • Keep it short and dignified -- do not argue or over-explain
  • Never blame the reviewer or question their experience

HIPAA Compliance (CRITICAL)

  • NEVER reference or confirm any medical conditions, diagnoses, treatments, medications, or health details -- even if the reviewer mentioned them in their review
  • NEVER confirm or deny that someone is or was a patient
  • Keep responses general: "your experience," "your visit," "your care" -- not "your diagnosis" or "your treatment"
  • If a reviewer shares health details in their review, do NOT reference those specifics in the reply. Respond to the sentiment and experience only.
  • When inviting someone to follow up, use "please contact our office" -- never suggest discussing their "case" or "medical records"
  • This applies to ALL star ratings, positive and negative

Things to Avoid

  • Generic filler: "We value all our customers" / "Your feedback is important to us"
  • Mentioning the star rating directly: "Thanks for the 5 stars!"
  • Being defensive about negative reviews
  • Making promises the business can't keep
  • Using the reviewer's full name unless they used it in their review
  • Emojis (unless the business brand is very casual and the operator approves it)
  • Referencing any health information, even if the patient shared it publicly (HIPAA)

Checking Pending Reviews

To see what's waiting for approval:

python3 ~/review-responder/gbp_reviews.py pending

Dependencies

  • Python 3 with: google-auth, google-auth-oauthlib, requests
  • Client config files in ~/review-responder/clients/
  • Telegram channel connected for approval messages
Usage Guidance
This package is coherent for managing Google Business reviews, but review the following before installing: - OAuth credentials and refresh tokens: The skill requires storing oauth_client_id, oauth_client_secret, and a long-lived refresh_token per client in plaintext files under the skill workspace (clients/*.json). Treat these as sensitive: restrict filesystem access, use correct permissions, and rotate/revoke tokens if compromised. - Telegram messaging: The scripts do NOT send Telegram messages themselves—you must have OpenClaw configured to send the drafted messages and to enforce the manual approval step. Verify your agent's heartbeat/automation will not auto-run the 'reply' command without human confirmation. - OAuth server: oauth_server.py is a convenience to capture refresh tokens remotely. If you expose it on a public VPS, run it under HTTPS and secure the server; by default the code sets OAUTHLIB_INSECURE_TRANSPORT to allow HTTP which is insecure. Prefer get_client_token.py run locally if you cannot secure a web server. - File writes: The skill creates pending/ and review_log.json in its workspace. Ensure the workspace location is appropriate and protected; do not place it in a path readable by untrusted users or bots. - Verify code yourself: This is a community-provided skill (source unknown). If you plan to run it in production, audit the Python code and consider storing tokens in a secret manager rather than plaintext files. If you trust the source and follow the operational cautions above (secure storage, HTTPS for onboarding server, agent-level checks to prevent auto-posting), the package is appropriate for its stated purpose.
Capability Analysis
Type: OpenClaw Skill Name: google-review-responder Version: 1.0.0 The skill bundle is a legitimate tool designed to manage Google Business Profile reviews through an OpenClaw agent with a human-in-the-loop approval process. It includes Python scripts for OAuth2 authentication and API interaction (gbp_reviews.py, oauth_server.py), and the agent instructions (SKILL.md) provide detailed guidelines for professional responses and strict HIPAA compliance. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found; the code follows standard integration patterns for Google APIs.
Capability Tags
requires-oauth-token
Capability Assessment
Purpose & Capability
The name/description (monitor GBP reviews, draft replies, and send via Telegram for approval) matches the code: gbp_reviews.py polls Google APIs and can post replies. The skill legitimately needs OAuth client credentials and refresh tokens saved per-client. One mismatch: there is no built-in Telegram integration in the code—SKILL.md expects the OpenClaw agent to send the draft messages via an existing Telegram channel. This is a design choice (agent handles messaging) but worth noting.
Instruction Scope
SKILL.md instructs the agent to run the included scripts, save pending review JSON files, send drafts via Telegram, and only post after operator approval. The code only accesses client config files and Google APIs and writes pending and log files in its workspace; it does not read unrelated system files or exfiltrate data to unknown endpoints. HIPAA rules are explicitly included in the guidelines.
Install Mechanism
No install spec is provided (instruction-only skill). The included Python scripts declare standard pip dependencies (google-auth, google-auth-oauthlib, requests, Flask optionally). There are no downloads from untrusted URLs or unusual install steps in the package.
Credentials
The skill does not request environment variables but does require per-client OAuth credentials (oauth_client_id, oauth_client_secret, and long-lived refresh_token) stored in clients/*.json. Those credentials are sensitive but are necessary for the stated purpose. Ensure you understand that refresh tokens allow ongoing API access until revoked and are saved in plaintext files by the scripts.
Persistence & Privilege
Skill does not set always:true and does not alter other skills. It will create files in its workspace (clients/, pending/, review_log.json) which is normal for its function. Note: the platform's autonomous invocation (default) could run the 'check' flow on heartbeat; the skill and scripts rely on operator approval before calling the reply flow—if an agent or operator misconfigures automation to call the reply command without approval, replies could be posted. That is operational risk, not intrinsic malicious behavior in the skill.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install google-review-responder
  3. After installation, invoke the skill by name or use /google-review-responder
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Google Business Review Responder 1.0.0 – Initial Release - Monitors Google Business Profile reviews for multiple clients. - Drafts professional, compliant responses based on review content and guidelines. - Sends drafted replies to operators via Telegram for approval before posting. - Waits for operator approval, edits, or skip instructions before responding publicly. - Strong HIPAA compliance rules to protect client and patient information. - Supports review management through command line interface.
Metadata
Slug google-review-responder
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Google Business Review Responder?

Monitors new Google Business reviews, drafts tailored responses sent via Telegram for operator approval before posting publicly. It is an AI Agent Skill for Claude Code / OpenClaw, with 98 downloads so far.

How do I install Google Business Review Responder?

Run "/install google-review-responder" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Google Business Review Responder free?

Yes, Google Business Review Responder is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Google Business Review Responder support?

Google Business Review Responder is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Google Business Review Responder?

It is built and maintained by chris-openclaw (@chris-openclaw); the current version is v1.0.0.

💬 Comments