customer-pulse-check
anthropics/knowledge-work-plugins
Synthesize customer feedback from PayPal, HubSpot, and reviews into actionable issues with response templates.
What is customer-pulse-check?
Aggregates dispute, ticket, and review data across your connected platforms to identify the top 3 fixable customer issues. Clusters feedback into themes, rates them by frequency and impact, and generates draft response templates for owner review—with no automatic actions.
- Pulls PayPal disputes, HubSpot tickets, and review exports for a configurable lookback period
- Clusters feedback into themes and classifies by type (quality, delivery, billing, communication, expectation mismatch)
- Rates each theme by frequency and impact (high/medium/low revenue risk)
- Selects top 3 fixable issues with root-cause analysis and operational fix suggestions
- Drafts customer response templates ready for owner review and approval
How to install customer-pulse-check
npx skills add https://github.com/anthropics/knowledge-work-plugins --skill customer-pulse-check- At least one connected source: PayPal, HubSpot, or a review export CSV file
- Optional: review export files (Google Reviews CSV, Yelp export, etc.) uploaded to Files
How to use customer-pulse-check
- 1.Run the skill with optional `--since YYYY-MM-DD` argument (defaults to last 30 days)
- 2.Review the summary table showing total signals and top 3 issues with frequency and impact ratings
- 3.Read the draft response templates for each issue
- 4.Select which templates you want to send
- 5.Approve each template before it is sent to customers
Use cases
- Weekly customer health check: identify recurring complaints before they escalate to churn
- Post-launch review: synthesize early customer feedback to spot product or delivery gaps
- Billing dispute spike: cluster payment-related complaints to find systemic issues
- Support ticket triage: surface the most impactful themes to prioritize team effort
- Review management: respond to common themes across Google Reviews, Yelp, and other exports
- Customer success managers and support leads
- Product managers investigating customer pain points
- Business owners doing weekly health checks
- Teams managing multiple feedback channels
customer-pulse-check FAQ
The skill degrades gracefully. It will pull from whatever sources are connected (PayPal, HubSpot, or reviews) and skip missing ones. If no sources are connected, it will ask you to connect at least one.
No. The skill only drafts templates and presents them for your review. You must explicitly approve each response before it is sent, and no tickets or disputes are resolved without your confirmation.
The default lookback window is the last 30 days. You can customize this with the `--since YYYY-MM-DD` argument.
Issues are ranked by frequency (how many signals mention them) multiplied by impact rating (high/medium/low based on revenue risk and churn potential).
The summary table uses only first name and last initial to protect customer PII. Full details appear only in the draft response templates for your review.
Full instructions (SKILL.md)
Source of truth, from anthropics/knowledge-work-plugins.
name: customer-pulse-check description: Synthesizes themes from PayPal disputes, HubSpot tickets, and review exports into a top-3 fixable issues list with drafted response templates. Accepts optional since-date argument. allowed-tools: Read, WebFetch, Bash
Run the customer voice synthesis. Pull feedback signals from all connected sources, identify the themes that are actually fixable, and produce drafted responses the owner can review and send.
Parse arguments:
--since(default: last 30 days) — start dateYYYY-MM-DDfor the lookback window
Step 1 — Gather feedback signals
Using the customer-pulse skill workflow:
- Pull PayPal disputes and chargebacks for the period: reason codes, amounts, resolution status.
- Pull HubSpot support tickets and conversation notes for the period.
- If review export files are available (Google Reviews CSV, Yelp export, etc.) in Files: read and parse them.
- Count total signals per source.
Step 2 — Theme extraction
Cluster all signals into recurring themes. For each theme:
- Count how many signals mention it
- Classify: Product quality / Delivery / Billing / Communication / Expectation mismatch / Other
- Rate impact: 🔴 High (revenue risk, churn) / 🟡 Medium / 🟢 Low
Step 3 — Top-3 fixable issues
Using the ticket-deflector skill workflow:
Select the top 3 themes by: frequency × impact rating. For each:
- State the issue in one sentence
- Explain the root cause (where evident)
- Suggest a specific operational fix
- Draft a customer response template
Response template format:
Subject: Re: {issue topic}
Hi {first name},
Thank you for reaching out. {Acknowledgment of their experience in 1-2 sentences}.
{What we're doing about it / what happened / resolution offered}.
{Next step or offer}.
{Sign-off}
Step 4 — Summary table
Format the output as:
Customer Voice — {date range}
Total signals: {n} ({PayPal disputes: n} | {HubSpot tickets: n} | {Reviews: n})
TOP 3 FIXABLE ISSUES
1. {Issue} ({frequency}) — {impact} — Fix: {one-line fix}
2. {Issue} ({frequency}) — {impact} — Fix: {one-line fix}
3. {Issue} ({frequency}) — {impact} — Fix: {one-line fix}
Connector failures
Run with whatever sources are connected — this command degrades gracefully. If PayPal is missing, skip dispute data and note "PayPal not connected — dispute data skipped." If HubSpot is missing, skip ticket data and note it. If no sources are connected at all, stop and tell the owner: "No feedback sources connected. Connect at least one of PayPal, HubSpot, or upload a review export CSV."
Approval gates
- Never send response emails automatically. Present drafts for owner review only.
- Never close HubSpot tickets or resolve PayPal disputes without explicit owner confirmation.
- Never include customer PII in the summary — use first name + last initial only.
Output
Present the summary table, then each response template. Ask the owner which templates they'd like to send, then wait for explicit approval before drafting the send.
Related skills
More from anthropics/knowledge-work-plugins and the wider catalog.

customer-research
Multi-source research on customer questions with source attribution and confidence scoring.

daily-briefing
Morning rundown of meetings, closing deals, and waiting emails with account context.

data-context-extractor
Extract tribal data knowledge from analysts and generate company-specific data analysis skills.

data-visualization
Pick the right chart and generate publication-quality Python visualizations with built-in design and accessibility best practices.

debug
Structured debugging session to reproduce, isolate, diagnose, and fix issues systematically.

debug-zoom
Debug Zoom integration failures by isolating the layer and ranking hypotheses with verification steps.