🎁 Get the FREE AI Skills Starter Guide β€” Subscribe β†’
BytesAgainBytesAgain
πŸ¦€ ClawHub

Sentiment Monitor

by @wubinghuang2019-source

Public opinion monitoring and structured report generation. Collects publicly available web information about celebrities, TV shows, or brands, performs sent...

Versionv0.4.0
Downloads689
Stars⭐ 2
TERMINAL
clawhub install sentiment-monitor

πŸ“– About This Skill


name: sentiment-monitor version: "0.3.1" license: MIT description: > Public opinion monitoring and structured report generation. Collects publicly available web information about celebrities, TV shows, or brands, performs sentiment analysis, and generates responsive HTML reports. Trigger keywords: sentiment, public opinion, reputation, monitoring report. Examples: "monitor reputation of X", "public opinion analysis for X".

Public Opinion Monitoring & Report Generation

This skill provides a structured methodology for monitoring and analyzing public sentiment about public-facing entities (celebrities, entertainment products, brands) using publicly available web information.

Purpose & Scope

  • Input: A target entity name and optional time range
  • Output: A self-contained HTML report with sentiment analysis, event timeline, and actionable recommendations
  • Data sources: Public web pages, news articles, social media (public content only)
  • Output format: HTML file with embedded CSS (no external dependencies)
  • Prerequisites

    This skill requires the AI assistant platform to provide the following built-in tools:

    | Tool | Role | Required | |------|------|----------| | WebSearch | Retrieve public web information | Yes | | WebFetch | Read detailed web page content | Yes | | Write | Save the generated HTML report | Yes | | preview_url | Display the HTML report preview | Optional | | deliver_attachments | Deliver the report file to the user | Optional |

    > Note: If the platform does not provide WebSearch or WebFetch, this skill cannot function. > In that case, inform the user that the required web access tools are not available.

    This skill has zero external dependencies. It does not install packages, execute scripts, access system files, or communicate with external servers.

    Privacy & Data Policy

  • All data comes from publicly accessible web sources only
  • No private data, personal information, or non-public content is accessed
  • Aggregate statistics (e.g., view counts, play counts) are only collected when explicitly displayed on public pages
  • Missing data is labeled as "Data Unavailable" in reports β€” no data is fabricated or estimated
  • Reports are saved to the user's local workspace; no data leaves the local environment
  • Data Availability Caveats

    > Important context for agents: > > The skill instructs gathering platform metrics (social media trending, view counts, engagement data). > In practice, most specific numerical metrics (e.g., exact view counts, follower numbers) are only accessible > through platform-specific APIs and are typically not available via public web search or page fetching. > > The agent should: > - Only report metrics that are explicitly visible in the web search results or fetched page content > - Never estimate, extrapolate, or fabricate numerical data > - Use qualitative descriptions (e.g., "high activity", "significant engagement") when exact numbers are unavailable > - Label any unavailable data as "Data Unavailable"

    Trigger Conditions

    This skill activates when user requests include keywords related to:

  • Public opinion, sentiment, reputation monitoring
  • Celebrity or brand image analysis
  • Media coverage or buzz tracking
  • Supported Monitoring Targets

  • Entertainment figures: Actors, musicians, influencers
  • Media products: TV series, films, variety shows
  • Brands and products: Consumer brands, product launches
  • Workflow

    Phase 1: Parameter Identification

    Extract from the user request:

    | Parameter | Default | Description | |-----------|---------|-------------| | Target | User-specified | Name of the entity to monitor | | Time range | Last 30 days | Customizable by user | | Output format | HTML | Only HTML format is supported | | Focus areas | All dimensions | Fan data, reputation events, platform activity, risk alerts |

    Phase 2: Public Information Gathering

    The AI assistant gathers information from publicly accessible web sources across these dimensions:

    Basic data (3-5 searches recommended):

  • Recent news coverage
  • Trending topics and hot searches
  • Public reviews and discussions
  • Fan community activity
  • Cross-platform discussion volume
  • Event identification:

  • Positive events: product releases, awards, charity work
  • Negative events: controversies, criticism, backlash
  • For each event: timestamp, platform, sentiment direction
  • > Important: Only data obtainable through public web searches should be used. > If specific numerical data is unavailable, the report should indicate "Data Unavailable".

    Phase 3: Analysis

    3.1 Sentiment Classification Categorize the overall public opinion into:

  • Positive proportion (favorable coverage, support, enthusiasm)
  • Neutral proportion (objective reporting, general discussion)
  • Negative proportion (criticism, controversy, backlash)
  • 3.2 Event Timeline Organize major events chronologically (minimum 5 events), including:

  • Event name and description
  • Timestamp
  • Sentiment direction (positive / neutral / negative)
  • Impact level (high / medium / low)
  • 3.3 Risk Assessment Identify current or potential reputation risks:

  • Active negative sentiment trends
  • Sensitive topics that may escalate
  • Competitive disadvantages
  • Phase 4: Report Generation

    Generate a complete, self-contained HTML report using the template structure below. The template is provided inline β€” no external file dependencies are required.

    File naming: {target_name}{days}day_sentiment_report_{YYYYMMDD}.html

    Report sections:

    1. Header β€” Title, monitoring period, generation date 2. KPI Overview β€” Overall sentiment score (0-100), positive/neutral/negative ratios, event count, active platforms 3. Sentiment Distribution β€” Visual bar chart and trend description 4. Event Timeline β€” Chronological list with sentiment and impact tags 5. Platform Performance β€” Cross-platform activity comparison 6. Risk Assessment Matrix β€” Current risks with severity and recommendations 7. Operational Recommendations β€” Short-term (1-2 weeks), mid-term (1 month), long-term (3+ months) 8. Disclaimer β€” Data source clarification

    Phase 5: Delivery

    1. Save the HTML report using the Write tool 2. Display the report preview using preview_url (if available) 3. Deliver the file using deliver_attachments (if available) 4. If tools are unavailable, provide the file path for manual access

    Guidelines

  • All data must originate from public web sources β€” fabrication is strictly prohibited
  • Unavailable data must be clearly marked as "Data Unavailable"
  • Time ranges are calculated automatically based on the current date
  • Sentiment analysis should be evidence-based, avoiding subjective assumptions
  • Reports should maintain objectivity, presenting multiple viewpoints for controversial events
  • Sensitive political content should be avoided in reports
  • Report Template

    Use the following HTML structure as the report template. All CSS is embedded inline β€” no external stylesheets, fonts, or scripts are required. Replace {PLACEHOLDER} tokens with actual data.

    
    
    
        
        
        {TITLE}
        
    
    
        

    {TITLE}

    {PERIOD} {DATE}
    {POSITIVE}%
    Positive
    {NEUTRAL}%
    Neutral
    {NEGATIVE}%
    Negative
    {HOT_COUNT}
    Key Events
    📊 Sentiment Distribution
    Positive {POSITIVE}% Neutral {NEUTRAL}% Negative {NEGATIVE}%

    {TREND_SUMMARY}

    📅 Key Events Timeline
    {EVENT_DATE}
    {EVENT_TITLE}
    {EVENT_DESC}
    {EVENT_SENTIMENT} {EVENT_IMPACT_LABEL}
    📰 Platform Performance
    Risk Assessment
    RiskLevelRecommendation
    {RISK_DESC}{RISK_LEVEL_LABEL}{RISK_SUGGESTION}
    💡 Recommendations

    Short-term (1-2 weeks)

    • {ADVICE_SHORT_1}
    • {ADVICE_SHORT_2}

    Mid-term (1 month)

    • {ADVICE_MID_1}
    • {ADVICE_MID_2}

    Long-term (3+ months)

    • {ADVICE_LONG_1}
    • {ADVICE_LONG_2}

    This report is based on publicly available web information and is for reference only.

    Report generated by AI; does not represent any official position.

    βš™οΈ Configuration

    This skill requires the AI assistant platform to provide the following built-in tools:

    | Tool | Role | Required | |------|------|----------| | WebSearch | Retrieve public web information | Yes | | WebFetch | Read detailed web page content | Yes | | Write | Save the generated HTML report | Yes | | preview_url | Display the HTML report preview | Optional | | deliver_attachments | Deliver the report file to the user | Optional |

    > Note: If the platform does not provide WebSearch or WebFetch, this skill cannot function. > In that case, inform the user that the required web access tools are not available.

    This skill has zero external dependencies. It does not install packages, execute scripts, access system files, or communicate with external servers.