Listen Skill

Aggregate customer feedback from multiple channels with automatic source discovery.

When to Use

•Starting product requirements (gather customer needs)
•Investigating customer pain points
•Reviewing recent feedback
•User asks to "listen to customers" or "what are customers saying"

Capability Discovery

The skill automatically discovers available feedback capabilities:

support-tickets capability

bash

# Discovers CLI tools, APIs, and exports that provide ticket data
# Examples: ticket system CLIs, support platform exports, help desk APIs

customer-feedback capability

bash

# Discovers feedback platforms, voting systems, feature request tools
# Examples: feedback exports, product board files, customer labels in issue trackers

surveys capability

bash

# Discovers survey response exports (CSV, JSON)
# Examples: survey platform exports, NPS data files, form responses

conversations capability

bash

# Discovers chat and messaging data sources
# Examples: chat exports, messaging platform data

issue-tracking capability

bash

# Discovers bug/issue tracking systems with customer-reported items
# Examples: issue tracker CLIs with customer labels or tags

Usage

bash

# Listen to all available capabilities (auto-discovers sources)
/wicked-product:listen

# Specific capability type
/wicked-product:listen --capability support-tickets

# Time window
/wicked-product:listen --days 30
/wicked-product:listen --since "2026-01-01"

# Filter by tags
/wicked-product:listen --tags bug,feature-request

# Limit results
/wicked-product:listen --limit 50

Aggregation Strategy

•
Discover Available Capabilities
- •Check for support-tickets capability (CLI tools, APIs)
- •Check for customer-feedback capability (exports, voting systems)
- •Check for surveys capability (CSV/JSON exports)
- •Check for conversations capability (chat/messaging data)
- •Check for issue-tracking capability (customer-labeled issues)
- •Look for saved feedback in voice store
•
Fetch Recent Feedback
- •Default: Last 30 days
- •Configurable timeframe
- •Filter by source, tag, or keyword
•
Normalize Format
- •Extract: ID, source, date, author, content, sentiment
- •Tag with categories (bug, feature, praise, complaint)
- •Store in unified format

•

Store for Analysis

code

~/.something-wicked/voice/feedback/{source}/{date}/{id}.md

Unified Feedback Format

yaml

---
id: fb_abc123
capability: support-tickets
source_tool: detected-cli-tool
source_id: ticket_456
date: 2026-01-20T10:30:00Z
author: customer@example.com
segment: enterprise
tags: [bug, mobile, performance]
sentiment: negative
priority: high
---

# {Title/Subject}

## Original Feedback
{Raw customer quote}

## Context
- Platform: iOS 16
- Plan: Enterprise
- User since: 2025-06-01

Output

Concise summary of aggregated feedback:

markdown

## Listening Report: {Timeframe}

### Capabilities Discovered
- support-tickets: {X} items from {N} source(s)
- customer-feedback: {Y} items from {N} source(s)
- surveys: {Z} responses from {N} source(s)
- Total: {N} feedback items across {M} capabilities

### Quick Stats
- Sentiment: {%pos} positive, {%neg} negative, {%neu} neutral
- Top tags: {tag1} ({count}), {tag2} ({count})
- Critical items: {count}

### Recent Highlights
1. **{Title}** - {source} - {date}
   - Sentiment: {sentiment}
   - Quote: "{excerpt}"

{Top 5-10 items}

### Next Steps
Run /wicked-product:analyze to extract themes and trends.

See channels.md for detailed capability integration patterns.

Integration

With wicked-mem

python

# Recall past customer insights
if has_plugin("wicked-mem"):
    memories = recall(f"customer feedback about {topic}")
    # Provide historical context

With wicked-crew

python

# Auto-trigger during product:requirements phase
if event == "product:requirements:started":
    feedback = listen(days=30, tags=["feature-request"])
    inject_context(feedback)

Storage

Feedback stored at: ~/.something-wicked/voice/feedback/{source}/{YYYY-MM}/{id}.md

Error Handling

•Source unavailable: Skip gracefully, report in summary
•No recent feedback: Report "No feedback found in timeframe"
•API rate limits: Cache and resume
•Invalid format: Log warning, continue with valid items

Rules

•Never expose customer PII beyond what's already in feedback
•Respect source rate limits
•Cache API responses when possible
•Report source availability in output
•Keep summaries concise (top 10 items max)