legislative-monitoring

name	legislative-monitoring
description	Voting pattern analysis, committee effectiveness, bill tracking, parliamentary oversight for Swedish intelligence operations
license	Apache-2.0

Legislative Monitoring Skill

🔴 AI FIRST Quality Principle

This skill MUST be applied with the AI FIRST principle: never accept first-pass quality. ALL analysis and content MUST go through minimum 2 complete iterations. After first pass, read ALL output back completely and systematically improve every section — strengthen evidence, deepen analysis, add specific citations, broaden perspectives. Spend ALL allocated time on real work. Single-pass output is NEVER acceptable. NO SHORTCUTS.

Purpose

This skill provides comprehensive legislative monitoring capabilities for the Riksdagsmonitor platform, enabling systematic tracking of parliamentary activities, voting patterns, committee effectiveness, bill progression, and oversight mechanisms. It transforms raw parliamentary data into actionable intelligence for understanding legislative dynamics, predicting policy outcomes, and assessing institutional effectiveness within the Swedish Riksdag.

When to Use This Skill

Apply this skill when:

• ✅ Tracking bill progression through legislative process
• ✅ Analyzing voting deviation patterns within and across parties
• ✅ Assessing committee productivity and effectiveness
• ✅ Monitoring parliamentary oversight activities (questions, interpellations)
• ✅ Predicting legislative outcomes based on historical patterns
• ✅ Identifying cross-party collaboration opportunities
• ✅ Evaluating individual parliamentarian productivity
• ✅ Detecting legislative obstruction or procedural manipulation
• ✅ Measuring government vs. opposition effectiveness
• ✅ Tracking amendment success rates and strategic positioning
• ✅ Tracking Riksrevisionen (RiR) audit findings and constitutional skrivelse deadlines
• ✅ Monitoring government accountability for audit recommendations
• ✅ Generating overdue-deadline alerts for unanswered RiR reports

Do NOT use for:

• ❌ Manipulating legislative processes through intelligence
• ❌ Interfering with democratic parliamentary procedures
• ❌ Undermining legitimate opposition activities
• ❌ Violating parliamentary privilege or confidentiality

Legislative Monitoring Framework

Five Core Analytical Dimensions

The Riksdagsmonitor platform monitors Swedish parliamentary activities across five critical dimensions to provide comprehensive legislative intelligence:

graph TB
    subgraph "Legislative Data Collection"
        A1[🗳️ Voting Records<br/>3.5M+ votes<br/>1971-present]
        A2[📄 Bill Tracking<br/>89K+ documents<br/>Full lifecycle]
        A3[👥 Committee Activity<br/>15 committees<br/>Productivity metrics]
        A4[🔍 Oversight Actions<br/>Questions, interpellations<br/>Accountability measures]
        A5[🤝 Cross-Party Collaboration<br/>Co-authorship networks<br/>Coalition patterns]
    end
    
    subgraph "Legislative Intelligence Analysis"
        A1 --> B1[Deviation Pattern Analysis]
        A2 --> B2[Bill Success Prediction]
        A3 --> B3[Committee Effectiveness Assessment]
        A4 --> B4[Oversight Effectiveness Tracking]
        A5 --> B5[Coalition Stability Measurement]
    end
    
    subgraph "Intelligence Products"
        B1 --> C1[Voting Intelligence]
        B2 --> C2[Legislative Forecasts]
        B3 --> C3[Institutional Health Reports]
        B4 --> C4[Accountability Metrics]
        B5 --> C5[Coalition Analysis]
    end
    
    subgraph "Strategic Intelligence"
        C1 & C2 & C3 & C4 & C5 --> D[📊 Legislative Dashboard]
        D --> E[Policy Outcome Predictions]
        D --> F[Institutional Risk Assessment]
        D --> G[Coalition Stability Forecasts]
    end
    
    style A1 fill:#e1f5ff
    style A2 fill:#e1f5ff
    style A3 fill:#e1f5ff
    style A4 fill:#e1f5ff
    style A5 fill:#e1f5ff
    style D fill:#ffe6cc
    style E fill:#ccffcc
    style F fill:#ffcccc
    style G fill:#fff9cc

1. Voting Deviation Pattern Analysis

Party Discipline Measurement

Voting deviation from party lines reveals internal party conflicts, ideological tensions, and coalition instability. The Riksdagsmonitor platform tracks these patterns with temporal granularity.

Database Views:

• view_riksdagen_vote_data_ballot_politician_summary_daily - Daily voting patterns
• view_riksdagen_vote_data_ballot_politician_summary_monthly - Monthly aggregates
• view_riksdagen_vote_data_ballot_politician_summary_annual - Annual trends
• view_riksdagen_politician_ballot_summary - Career-wide voting statistics
• view_riksdagen_politician_decision_pattern - Decision pattern classification

Deviation Intelligence Framework

@Service
public class VotingDeviationAnalyzer {
    
    /**
     * Analyzes voting deviation patterns to understand party discipline.
     * 
     * Intelligence applications:
     * - Identify factional splits within parties
     * - Predict coalition stability
     * - Assess leadership control
     * - Detect ideological realignment
     * - Forecast party breakaway risks
     */
    
    public DeviationIntelligence analyzePartyDeviation(
        String partyId, 
        LocalDate startDate, 
        LocalDate endDate
    ) {
        String sql = """
            WITH party_voting AS (
                SELECT 
                    v.ballot_id,
                    b.vote_date,
                    b.issue_category,
                    b.issue_subcategory,
                    b.vote_outcome,
                    
                    -- Party position (mode of party votes)
                    MODE() WITHIN GROUP (ORDER BY v.vote) as party_position,
                    
                    -- Deviation metrics
                    COUNT(*) as party_member_votes,
                    COUNT(DISTINCT v.vote) as vote_diversity,
                    
                    -- Rebel identification
                    COUNT(CASE 
                        WHEN v.vote != (MODE() WITHIN GROUP (ORDER BY v.vote)) 
                        THEN 1 END
                    ) as rebel_count,
                    
                    -- Rebel percentage
                    COUNT(CASE 
                        WHEN v.vote != (MODE() WITHIN GROUP (ORDER BY v.vote)) 
                        THEN 1 END
                    )::float / NULLIF(COUNT(*), 0) * 100 as rebel_percentage,
                    
                    -- Identify specific rebels
                    ARRAY_AGG(
                        CASE 
                            WHEN v.vote != (MODE() WITHIN GROUP (ORDER BY v.vote))
                            THEN p.person_id::text
                        END
                    ) FILTER (WHERE v.vote != (MODE() WITHIN GROUP (ORDER BY v.vote))) as rebel_ids
                    
                FROM vote v
                JOIN ballot b ON v.ballot_id = b.ballot_id
                JOIN person p ON v.person_id = p.person_id
                WHERE p.party = :partyId
                    AND b.vote_date BETWEEN :startDate AND :endDate
                GROUP BY v.ballot_id, b.vote_date, b.issue_category, 
                         b.issue_subcategory, b.vote_outcome
            ),
            deviation_patterns AS (
                SELECT 
                    issue_category,
                    COUNT(*) as total_votes,
                    AVG(rebel_percentage) as avg_deviation_rate,
                    MAX(rebel_percentage) as max_deviation_rate,
                    STDDEV(rebel_percentage) as deviation_volatility,
                    
                    -- High-deviation votes (>15% rebels)
                    COUNT(CASE WHEN rebel_percentage > 15 THEN 1 END) as high_deviation_votes,
                    
                    -- Identify consistent rebels
                    (
                        SELECT person_id, COUNT(*) as rebel_count
                        FROM UNNEST(ARRAY_AGG(rebel_ids)) as person_id
                        GROUP BY person_id
                        ORDER BY COUNT(*) DESC
                        LIMIT 10
                    ) as top_rebels
                    
                FROM party_voting
                GROUP BY issue_category
            )
            SELECT 
                p.party as party_id,
                p.party_name,
                
                -- Overall deviation metrics
                AVG(dp.avg_deviation_rate) as overall_deviation_rate,
                MAX(dp.max_deviation_rate) as peak_deviation,
                AVG(dp.deviation_volatility) as avg_volatility,
                
                -- High-risk categories (high deviation)
                ARRAY_AGG(
                    dp.issue_category 
                    ORDER BY dp.avg_deviation_rate DESC
                ) FILTER (WHERE dp.avg_deviation_rate > 10) as high_risk_categories,
                
                -- Discipline classification
                CASE 
                    WHEN AVG(dp.avg_deviation_rate) < 3 THEN 'HIGHLY_DISCIPLINED'
                    WHEN AVG(dp.avg_deviation_rate) < 7 THEN 'MODERATELY_DISCIPLINED'
                    WHEN AVG(dp.avg_deviation_rate) < 12 THEN 'LOW_DISCIPLINE'
                    ELSE 'FRACTURED'
                END as discipline_classification,
                
                -- Risk assessment
                CASE 
                    WHEN AVG(dp.avg_deviation_rate) > 15 
                        OR MAX(dp.max_deviation_rate) > 30 
                        THEN 'CRITICAL_INSTABILITY'
                    WHEN AVG(dp.avg_deviation_rate) > 10 
                        OR MAX(dp.max_deviation_rate) > 20 
                        THEN 'HIGH_FRACTURE_RISK'
                    WHEN AVG(dp.avg_deviation_rate) > 5 
                        THEN 'MODERATE_TENSION'
                    ELSE 'STABLE'
                END as stability_risk
                
            FROM party p
            JOIN deviation_patterns dp ON 1=1
            WHERE p.party = :partyId
            GROUP BY p.party, p.party_name
            """;
        
         jdbcTemplate.queryForObject(sql, DeviationIntelligence.class,
            Map.of(
                , partyId,
                , startDate,
                , endDate
            ));
    }
}

Deviation Pattern Taxonomy

Deviation Rate	Discipline Level	Political Implications	Intelligence Assessment
< 3%	HIGHLY_DISCIPLINED	Strong leadership control, unified ideology	Reliable coalition partner
3-7%	MODERATELY_DISCIPLINED	Occasional dissent, manageable tensions	Generally stable, monitor key issues
7-12%	LOW_DISCIPLINE	Frequent internal conflicts, weak leadership	Coalition risks, factional monitoring needed
> 12%	FRACTURED	Severe internal divisions, leadership crisis	High breakaway risk, unstable partner

Strategic Intelligence Applications

from typing import Dict, List
import pandas as pd
import numpy as np

class DeviationIntelligenceAnalyzer:
    """
    Applies deviation pattern intelligence to strategic assessments.
    
    Intelligence products:
    1. Coalition stability forecasts
    2. Leadership vulnerability assessment
    3. Policy area risk mapping
    4. Factional structure identification
    """
    
    def assess_coalition_stability(
        self, 
        coalition_parties: List[str], 
        date_from: str, 
        date_to: str
    ) -> Dict:
        """
        Assesses coalition stability through deviation pattern analysis.
        
        Coalition stability factors:
        - Intra-party discipline (low deviation = stable)
        - Cross-party voting alignment (high alignment = stable)
        - Deviation trend (increasing deviation = instability)
        """
        
        query = """
        WITH coalition_deviation AS (
            SELECT 
                p.party,
                b.vote_date,
                b.issue_category,
                
                -- Party deviation rate
                COUNT(CASE 
                    WHEN v.vote != (
                        SELECT MODE() WITHIN GROUP (ORDER BY v2.vote)
                        FROM vote v2
                        JOIN person p2 ON v2.person_id = p2.person_id
                        WHERE p2.party = p.party
                            AND v2.ballot_id = v.ballot_id
                    ) THEN 1 END
                )::float / NULLIF(COUNT(*), 0) * 100 as deviation_rate
                
            FROM vote v
            JOIN person p ON v.person_id = p.person_id
            JOIN ballot b ON v.ballot_id = b.ballot_id
            WHERE p.party = ANY(%s)
                AND b.vote_date BETWEEN %s AND %s
            GROUP BY p.party, b.vote_date, b.issue_category, v.ballot_id
        ),
        cross_party_alignment AS (
            -- Measure voting alignment between coalition parties
            SELECT 
                p1.party as party_a,
                p2.party as party_b,
                COUNT(CASE WHEN v1.vote = v2.vote THEN 1 END)::float / 
                NULLIF(COUNT(*), 0) * 100 as alignment_rate
            FROM vote v1
            JOIN person p1 ON v1.person_id = p1.person_id
            JOIN vote v2 ON v1.ballot_id = v2.ballot_id
            JOIN person p2 ON v2.person_id = p2.person_id
            WHERE p1.party = ANY(%s)
                AND p2.party = ANY(%s)
                AND p1.party < p2.party  -- Avoid duplicates
            GROUP BY p1.party, p2.party
        )
        SELECT 
            cd.party,
            AVG(cd.deviation_rate) as avg_deviation,
            STDDEV(cd.deviation_rate) as deviation_volatility,
            
            -- Trend analysis (linear regression slope)
            REGR_SLOPE(
                cd.deviation_rate, 
                EXTRACT(EPOCH FROM cd.vote_date)
            ) as deviation_trend,
            
            -- Cross-party alignment (average with coalition partners)
            (SELECT AVG(cpa.alignment_rate) 
             FROM cross_party_alignment cpa 
             WHERE cpa.party_a = cd.party OR cpa.party_b = cd.party
            ) as coalition_alignment
            
        FROM coalition_deviation cd
        GROUP BY cd.party
        """
        
        results = pd.read_sql(
            query, 
            .connection, 
            params=[coalition_parties, date_from, date_to, 
                   coalition_parties, coalition_parties]
        )
        
        
        stability_components = []
        
         _, party  results.iterrows():
            
            deviation_stability = (, ( - party[]) / ) * 
            
            
            alignment_stability = (party[] / ) * 
            
            
            trend_stability = (,  - (party[]) * )
            
            party_stability = deviation_stability + alignment_stability + trend_stability
            stability_components.append({
                : party[],
                : (party_stability, ),
                : (party[], ),
                : (party[], ),
                : (party[], )
            })
        
        
        overall_stability = np.mean([p[]  p  stability_components])
        
         {
            : coalition_parties,
            : ,
            : (overall_stability, ),
            : stability_components,
            : ._classify_stability(overall_stability),
            : ._assess_stability_risks(stability_components)
        }
    
     () -> :
        
         score >= :
             
         score >= :
             
         score >= :
             
        :
             
    
     () -> []:
        
        risks = []
        
         party  components:
             party[] > :
                risks.append()
             party[] < :
                risks.append()
             party[] > :
                risks.append()
        
         risks  risks  []

2. Committee Effectiveness Assessment

Parliamentary Committee Intelligence

Committees are the workhorses of legislative bodies. The Riksdagsmonitor platform measures their productivity, influence, and effectiveness.

Database Views:

• view_riksdagen_committee_decisions - Committee decision tracking
• view_riksdagen_committee_document - Committee-produced documents
• view_riksdagen_committee_member_activity - Individual member contributions
• view_committee_effectiveness_metrics - Composite effectiveness measures

Committee Productivity Metrics

-- Comprehensive Committee Effectiveness Assessment
WITH committee_activity AS (
    SELECT 
        c.org_code,
        c.committee_name,
        c.policy_area,
        
        -- Document production
        COUNT(DISTINCT cd.document_id) as total_documents,
        COUNT(DISTINCT CASE WHEN cd.document_type = 'committee_report' 
                           THEN cd.document_id END) as reports_produced,
        COUNT(DISTINCT CASE WHEN cd.document_type = 'committee_proposal' 
                           THEN cd.document_id END) as proposals_made,
        
        -- Legislative impact
        COUNT(DISTINCT CASE WHEN cd.adopted_by_parliament = TRUE 
                           THEN cd.document_id END) as adopted_proposals,
        COUNT(DISTINCT CASE WHEN cd.influenced_final_legislation = TRUE 
                           THEN cd.document_id END) as legislative_influence,
        
        -- Member engagement
        COUNT( cm.person_id)  member_count,
        (cm.attendance_rate)  avg_attendance,
        (cm.contribution_score)  avg_contribution,
        
        
        (   cd.document_type   
                            cd.document_id )  oversight_actions,
        (   cd.oversight_response_received   
                            cd.document_id )  oversight_responses,
        
        
        (cd.processing_time_days)  avg_processing_time,
        STDDEV(cd.processing_time_days)  processing_time_variance,
        
        
        (
             ( dp.person_id)::  
                   ( p.party)
             document_person dp
             person p  dp.person_id  p.person_id
             dp.document_id  (
                 document_id 
                 committee_document 
                 org_code  c.org_code
            )
        )  cross_party_engagement
        
     committee c
      committee_document cd  c.org_code  cd.org_code
      committee_member cm  c.org_code  cm.org_code
     c.active  
         cd.created_date  NOW()   
      c.org_code, c.committee_name, c.policy_area
),
committee_effectiveness  (
     
        ca.,
        
        
        (
            LEAST(ca.total_documents  , )     
            LEAST(ca.reports_produced  , )     
            (ca.adopted_proposals::  (ca.proposals_made, ))     
            (ca.oversight_responses::  (ca.oversight_actions, ))     
            ca.avg_attendance    
        )  productivity_score,
        
        
        (
            (ca.adopted_proposals::  (ca.proposals_made, ))   
            (ca.legislative_influence::  (ca.total_documents, ))   
            LEAST(ca.adopted_proposals  , )  
        )  impact_score,
        
        
        (
             
                 ca.avg_processing_time    
                 ca.avg_processing_time    
                 ca.avg_processing_time    
                 ca.avg_processing_time    
                 
            
        )  efficiency_score,
        
        
        (
            ca.cross_party_engagement   
            ca.avg_contribution  
        )  collaboration_score,
        
        
        (
            (
                LEAST(ca.total_documents  , )   
                LEAST(ca.reports_produced  , )   
                (ca.adopted_proposals::  (ca.proposals_made, ))   
                (ca.oversight_responses::  (ca.oversight_actions, ))   
                ca.avg_attendance  
            )     
            (
                (ca.adopted_proposals::  (ca.proposals_made, ))   
                (ca.legislative_influence::  (ca.total_documents, ))   
                LEAST(ca.adopted_proposals  , )  
            )     
            (
                 
                     ca.avg_processing_time    
                     ca.avg_processing_time    
                     ca.avg_processing_time    
                     ca.avg_processing_time    
                     
                
            )     
            (
                ca.cross_party_engagement   
                ca.avg_contribution  
            )    
        )  overall_effectiveness_score
        
     committee_activity ca
)
 
    org_code,
    committee_name,
    policy_area,
    total_documents,
    reports_produced,
    adopted_proposals,
    proposals_made,
    ROUND(productivity_score, )  productivity,
    ROUND(impact_score, )  impact,
    ROUND(efficiency_score, )  efficiency,
    ROUND(collaboration_score, )  collaboration,
    ROUND(overall_effectiveness_score, )  overall_effectiveness,
    
    
     
         overall_effectiveness_score    
         overall_effectiveness_score    
         overall_effectiveness_score    
         
      effectiveness_classification,
    
    
     
         overall_effectiveness_score   
             efficiency_score   
             collaboration_score   
             
         overall_effectiveness_score   
             efficiency_score   
             
         
      institutional_health
    
 committee_effectiveness
  overall_effectiveness_score ;

3. Bill Success Prediction

Legislative Outcome Forecasting

The Riksdagsmonitor platform predicts bill success based on historical patterns, sponsor characteristics, coalition dynamics, and timing factors.

from sklearn.ensemble import RandomForestClassifier
from sklearn.preprocessing import StandardScaler
import pandas as pd
import numpy as np

class BillSuccessPredictor:
    """
    Predicts bill passage likelihood using machine learning on historical data.
    
    Features:
    - Sponsor characteristics (party, experience, influence)
    - Coalition structure (government vs. opposition)
    - Bill characteristics (policy area, complexity, novelty)
    - Temporal factors (timing in parliamentary session)
    - Historical patterns (similar bill success rates)
    """
    
    def __init__(self):
        self.model = RandomForestClassifier(n_estimators=100, max_depth=10)
        self.scaler = StandardScaler()
        self.trained = False
    
    def prepare_training_data(self) -> Tuple[pd.DataFrame, pd.Series]:
        """
        Extracts historical bill data for model training.
        
        Returns features and labels (passed/failed).
        """
        
        query = """
        SELECT 
            d.document_id,
            d.document_type,
            d.issue_category,
            d.issue_subcategory,
            d.document_complexity_score,
            d.document_novelty_score,
            
            -- Sponsor characteristics
            p.party,
            p.years_in_parliament,
            vim.influence_score,
            vim.network_centrality,
            vim.collaboration_score,
            
            -- Coalition context
            CASE WHEN p.party IN (SELECT party FROM government_coalition) 
                 THEN TRUE ELSE FALSE END as is_government_sponsor,
            
            -- Bill characteristics
            LENGTH(dc.text_content) as bill_length,
            (SELECT COUNT(*) FROM document_person WHERE document_id = d.document_id) as co_sponsor_count,
            (SELECT COUNT(DISTINCT party) FROM document_person dp 
             JOIN person p2 ON dp.person_id = p2.person_id 
             WHERE dp.document_id = d.document_id) as party_diversity,
            
            -- Temporal factors
            EXTRACT(MONTH FROM d.created_date) as submission_month,
            EXTRACT(DAY FROM (
                SELECT MAX(s.session_end_date) FROM parliamentary_session s
                WHERE d.created_date BETWEEN s.session_start_date AND s.session_end_date
            ) - d.created_date) as days_before_session_end,
            
            -- Historical success rate for similar bills
            (
                SELECT AVG(CASE WHEN adopted = TRUE THEN 1.0 ELSE 0.0 END)
                FROM document d2
                WHERE d2.issue_category = d.issue_category
                    AND d2.created_date < d.created_date
                    AND d2.created_date >= d.created_date - INTERVAL '5 years'
            ) as category_historical_success_rate,
            
            -- Label: Did the bill pass?
            CASE WHEN d.adopted_by_parliament = TRUE THEN 1 ELSE 0 END as passed
            
        FROM document d
        JOIN document_person dp ON d.document_id = dp.document_id
        JOIN person p ON dp.person_id = p.person_id
        JOIN view_riksdagen_politician_influence_metrics vim ON p.person_id = vim.person_id
        JOIN document_content dc ON d.document_id = dc.document_id
        WHERE d.document_type IN ('motion', 'bill', 'proposal')
            AND d.final_status IS NOT NULL  -- Only completed bills
            AND d.created_date >= '2010-01-01'  -- Last 15 years
        """
        
        df = pd.read_sql(query, .connection)
        
        
        df[] = pd.Categorical(df[]).codes
        df[] = pd.Categorical(df[]).codes
        
        
        df[] = df[] * df[]
        df[] = df[].astype() * 
        df[] = np.where(df[] < , , )
        
        
        feature_cols = [
            , , ,
            , , ,
            , , ,
            , , , ,
            , ,
            , , 
        ]
        
        X = df[feature_cols]
        y = df[]
        
         X, y
    
     ():
        
        X, y = .prepare_training_data()
        
        
        X_scaled = .scaler.fit_transform(X)
        
        
        .model.fit(X_scaled, y)
        .trained = 
        
        
         sklearn.model_selection  cross_val_score
        cv_scores = cross_val_score(.model, X_scaled, y, cv=)
        
         {
            : (X),
            : (np.mean(cv_scores), ),
            : ((
                X.columns,
                [(imp, )  imp  .model.feature_importances_]
            ))
        }
    
     () -> :
        
        
          .trained:
             ValueError()
        
        
        
        
        
        probability = .model.predict_proba(bill_features_scaled)[][]
        
         {
            : bill_id,
            : (probability, ),
            :   probability >   
                           probability >   ,
            :   (probability - ) >   ,
            : ._identify_key_factors(bill_features),
            : ._generate_recommendations(
                bill_features, 
                probability
            )
        }
    
     () -> []:
        
        feature_importance = .model.feature_importances_
        top_features = (
            (features.columns, feature_importance, features.iloc[]),
            key= x: x[],
            reverse=
        )[:]
        
         [ 
                 feat, imp, val  top_features]
    
     () -> []:
        
        recommendations = []
        
         features[].iloc[] ==   probability < :
            recommendations.append(
                
            )
        
         features[].iloc[] < :
            recommendations.append(
                
            )
        
         features[].iloc[] < :
            recommendations.append(
                
            )
        
         features[].iloc[] < :
            recommendations.append(
                
            )
        
         features[].iloc[] > :
            recommendations.append(
                
            )
        
         recommendations  recommendations  [
            
        ]

4. Parliamentary Oversight Tracking

Accountability Mechanism Analysis

Parliamentary questions, interpellations, and oversight requests are key accountability tools. The Riksdagsmonitor platform tracks their effectiveness.

Database Views:

• view_riksdagen_oversight_activity - All oversight actions
• view_riksdagen_question_response_time - Response timeliness
• view_riksdagen_oversight_effectiveness - Impact assessment

@Service
public class OversightEffectivenessAnalyzer {
    
    /**
     * Analyzes parliamentary oversight effectiveness.
     * 
     * Metrics:
     * - Response rate (% answered)
     * - Response quality (substantive vs. evasive)
     * - Response timeliness (days to answer)
     * - Policy impact (changes resulting from oversight)
     * - Media attention (public accountability pressure)
     */
    
    public OversightProfile analyzeOversight(
        String politicianId, 
        LocalDate startDate, 
        LocalDate endDate
    ) {
        String sql = """
            WITH oversight_actions AS (
                SELECT 
                    d.document_id,
                    d.document_type,  -- 'question', 'interpellation', 'oversight_request'
                    d.created_date as submitted_date,
                    d.target_minister_id,
                    d.target_ministry,
                    d.issue_category,
                    
                    -- Response tracking
                    dr.response_document_id,
                    dr.response_date,
                    EXTRACT(DAY FROM dr.response_date - d.created_date) as response_time_days,
                    dr.response_quality_score,  -- 0-1, algorithmic assessment
                    dr.substantive_response,  -- Boolean: Did it address the question?
                    
                    -- Impact tracking
                    di.policy_change_resulted,
                    di.media_coverage_count,
                    di.follow_up_actions_count,
                    di.public_attention_score
                    
                FROM document d
                LEFT JOIN document_response dr ON d.document_id = dr.question_document_id
                LEFT JOIN document_impact di ON d.document_id = di.document_id
                JOIN document_person dp ON d.document_id = dp.document_id
                WHERE dp.person_id = :politicianId
                    AND d.document_type IN ('question', 'interpellation', 'oversight_request')
                    AND d.created_date BETWEEN :startDate AND :endDate
            )
            SELECT 
                p.person_id,
                p.first_name || ' ' || p.last_name as name,
                p.party,
                
                -- Volume metrics
                COUNT(*) as total_oversight_actions,
                COUNT(DISTINCT oa.issue_category) as issue_diversity,
                COUNT(DISTINCT oa.target_minister_id) as ministers_targeted,
                
                -- Response metrics
                COUNT(oa.response_document_id)::float / 
                NULLIF(COUNT(*), 0) * 100 as response_rate,
                AVG(oa.response_time_days) as avg_response_time,
                COUNT(CASE WHEN oa.substantive_response = TRUE THEN 1 END)::float /
                NULLIF(COUNT(oa.response_document_id), 0) * 100 as substantive_response_rate,
                AVG(oa.response_quality_score) * 100 as avg_response_quality,
                
                -- Impact metrics
                COUNT(CASE WHEN oa.policy_change_resulted = TRUE THEN 1 END) as policy_changes_achieved,
                AVG(oa.media_coverage_count) as avg_media_attention,
                SUM(oa.follow_up_actions_count) as total_follow_ups,
                AVG(oa.public_attention_score) * 100 as avg_public_attention,
                
                -- Effectiveness score (0-100)
                (
                    (COUNT(oa.response_document_id)::float / NULLIF(COUNT(*), 0) * 100) * 0.20 +  -- Response rate (20%)
                    (COUNT(CASE WHEN oa.substantive_response = TRUE THEN 1 END)::float /
                     NULLIF(COUNT(oa.response_document_id), 0) * 100) * 0.25 +  -- Substantive rate (25%)
                    (CASE 
                        WHEN AVG(oa.response_time_days) <= 7 THEN 100
                        WHEN AVG(oa.response_time_days) <= 14 THEN 75
                        WHEN AVG(oa.response_time_days) <= 30 THEN 50
                        WHEN AVG(oa.response_time_days) <= 60 THEN 25
                        ELSE 10
                    END) * 0.15 +  -- Timeliness (15%)
                    (COUNT(CASE WHEN oa.policy_change_resulted = TRUE THEN 1 END)::float /
                     NULLIF(COUNT(*), 0) * 100) * 0.25 +  -- Policy impact (25%)
                    (AVG(oa.public_attention_score) * 100) * 0.15  -- Public attention (15%)
                ) as oversight_effectiveness_score,
                
                -- Classification
                CASE 
                    WHEN (
                        (COUNT(oa.response_document_id)::float / NULLIF(COUNT(*), 0) * 100) * 0.20 +
                        (COUNT(CASE WHEN oa.substantive_response = TRUE THEN 1 END)::float /
                         NULLIF(COUNT(oa.response_document_id), 0) * 100) * 0.25 +
                        (CASE 
                            WHEN AVG(oa.response_time_days) <= 7 THEN 100
                            WHEN AVG(oa.response_time_days) <= 14 THEN 75
                            WHEN AVG(oa.response_time_days) <= 30 THEN 50
                            WHEN AVG(oa.response_time_days) <= 60 THEN 25
                            ELSE 10
                        END) * 0.15 +
                        (COUNT(CASE WHEN oa.policy_change_resulted = TRUE THEN 1 END)::float /
                         NULLIF(COUNT(*), 0) * 100) * 0.25 +
                        (AVG(oa.public_attention_score) * 100) * 0.15
                    ) >= 75 THEN 'HIGHLY_EFFECTIVE_OVERSIGHT'
                    WHEN (
                        (COUNT(oa.response_document_id)::float / NULLIF(COUNT(*), 0) * 100) * 0.20 +
                        (COUNT(CASE WHEN oa.substantive_response = TRUE THEN 1 END)::float /
                         NULLIF(COUNT(oa.response_document_id), 0) * 100) * 0.25 +
                        (CASE 
                            WHEN AVG(oa.response_time_days) <= 7 THEN 100
                            WHEN AVG(oa.response_time_days) <= 14 THEN 75
                            WHEN AVG(oa.response_time_days) <= 30 THEN 50
                            WHEN AVG(oa.response_time_days) <= 60 THEN 25
                            ELSE 10
                        END) * 0.15 +
                        (COUNT(CASE WHEN oa.policy_change_resulted = TRUE THEN 1 END)::float /
                         NULLIF(COUNT(*), 0) * 100) * 0.25 +
                        (AVG(oa.public_attention_score) * 100) * 0.15
                    ) >= 60 THEN 'MODERATELY_EFFECTIVE'
                    WHEN (
                        (COUNT(oa.response_document_id)::float / NULLIF(COUNT(*), 0) * 100) * 0.20 +
                        (COUNT(CASE WHEN oa.substantive_response = TRUE THEN 1 END)::float /
                         NULLIF(COUNT(oa.response_document_id), 0) * 100) * 0.25 +
                        (CASE 
                            WHEN AVG(oa.response_time_days) <= 7 THEN 100
                            WHEN AVG(oa.response_time_days) <= 14 THEN 75
                            WHEN AVG(oa.response_time_days) <= 30 THEN 50
                            WHEN AVG(oa.response_time_days) <= 60 THEN 25
                            ELSE 10
                        END) * 0.15 +
                        (COUNT(CASE WHEN oa.policy_change_resulted = TRUE THEN 1 END)::float /
                         NULLIF(COUNT(*), 0) * 100) * 0.25 +
                        (AVG(oa.public_attention_score) * 100) * 0.15
                    ) >= 40 THEN 'LOW_EFFECTIVENESS'
                    ELSE 'INEFFECTIVE_OVERSIGHT'
                END as effectiveness_classification
                
            FROM view_riksdagen_politician p
            JOIN oversight_actions oa ON 1=1
            WHERE p.person_id = :politicianId
            GROUP BY p.person_id, p.first_name, p.last_name, p.party
            """;
        
         jdbcTemplate.queryForObject(sql, OversightProfile.class,
            Map.of(
                , politicianId,
                , startDate,
                , endDate
            ));
    }
}

5. Cross-Party Collaboration Indicators

Coalition Building Intelligence

Cross-party collaboration signals potential coalition opportunities, ideological flexibility, and legislative compromise capacity.

Database Views:

• view_riksdagen_cross_party_collaboration - Co-authorship networks
• view_riksdagen_coalition_patterns - Historical coalition structures
• view_riksdagen_bipartisan_initiatives - Cross-party legislative efforts

-- Cross-Party Collaboration Network Analysis
WITH collaboration_events AS (
    SELECT 
        dp1.person_id as person_a,
        dp2.person_id as person_b,
        p1.party as party_a,
        p2.party as party_b,
        d.document_id,
        d.document_type,
        d.issue_category,
        d.created_date,
        d.adopted_by_parliament
    FROM document_person dp1
    JOIN document_person dp2 ON dp1.document_id = dp2.document_id
    JOIN person p1 ON dp1.person_id = p1.person_id
    JOIN person p2 ON dp2.person_id = p2.person_id
    JOIN document d ON dp1.document_id = d.document_id
    WHERE dp1.person_id < dp2.person_id  -- Avoid duplicates
        AND p1.party != p2.party  -- Cross-party only
        AND d.created_date >= NOW() - INTERVAL '4 years'
),
party_pair_collaboration AS (
    SELECT 
        ce.party_a,
        ce.party_b,
        COUNT(DISTINCT ce.document_id) as collaboration_count,
        COUNT(DISTINCT ce.person_a) + COUNT(DISTINCT ce.person_b) as unique_collaborators,
        COUNT(DISTINCT ce.issue_category)  issue_diversity,
        
        
        (  ce.adopted_by_parliament     ):: 
        ((), )    success_rate,
        
        
        (  ce.created_date  NOW()    
                     )  recent_collaborations,
        
        
        (
            ( ce.document_id)::  
            (( ce.person_a)  ( ce.person_b))   
            ( ce.issue_category)   
            ((  ce.adopted_by_parliament     ):: 
             ((), ))
        )  collaboration_intensity
        
     collaboration_events ce
      ce.party_a, ce.party_b
)
 
    party_a,
    party_b,
    collaboration_count,
    unique_collaborators,
    issue_diversity,
    ROUND(success_rate, )  success_rate,
    recent_collaborations,
    ROUND(collaboration_intensity, )  intensity_score,
    
    
     
         collaboration_intensity   
             success_rate   
             recent_collaborations   
             
         collaboration_intensity   
             recent_collaborations   
             
         collaboration_intensity   
             
         
      coalition_potential,
    
    
     
         issue_diversity    success_rate   
             
         collaboration_count   
             
         recent_collaborations  collaboration_count   
             
         
      relationship_type
    
 party_pair_collaboration
  intensity_score , collaboration_count ;

ISMS Compliance Mapping

ISO 27001:2022 Controls

Control	Legislative Monitoring Application
A.5.9 - Inventory of information and other associated assets	Catalog legislative data sources and integrity verification
A.8.16 - Monitoring activities	Continuous monitoring of parliamentary data feeds
A.8.24 - Use of cryptography	Secure storage of political intelligence data

NIST Cybersecurity Framework 2.0

Function	Legislative Monitoring Integration
IDENTIFY (ID.AM)	Asset inventory of legislative data sources
PROTECT (PR.DS)	Data integrity protection for parliamentary records
DETECT (DE.DP)	Detection of data manipulation or tampering

CIS Controls v8

Control	Application
CIS Control 8 - Audit Log Management	Legislative activity audit logging
CIS Control 11 - Data Recovery	Parliamentary data backup and recovery

7. Riksrevisionen (RiR) Follow-Up Tracking

Constitutional Framework

Swedish constitutional practice (Riksdagsordningen ch. 10 and Government Offices' established conventions) requires the government to formally respond to each Riksrevisionen audit report with a skrivelse (written communication to the Riksdag) within 4 calendar months of the report's publication.

Failure to respond within this deadline is a constitutional accountability gap that creates political risk and potential for opposition interpellations, motions, and media pressure.

Data Model

The authoritative dataset lives at data/rir-followups.json (JSON schema: schemas/rir-followups-schema.json). Each record captures:

Field	Description
rir_report_id	Riksdag document ID (e.g. HD01JuU31)
rir_number	Official RiR number (e.g. RiR 2026:6)
title / title_en	Swedish and English titles
agency	Primary audited government agency
committees	Riksdag committee codes (e.g. JuU, FöU)
publish_date	ISO 8601 publication date
skrivelse_deadline	Calculated 4-month constitutional deadline
gov_response_status	PENDING | RESPONDED | OVERDUE | PARTIAL
response_skrivelse_id	Reference if government has responded
parliamentary_followup_doc_ids	Related Riksdag documents
open_recommendations	Count of unresolved audit recommendations
risk_level	LOW | MEDIUM | HIGH | CRITICAL

Status Lifecycle

stateDiagram-v2
    [*] --> PENDING : RiR report published
    PENDING --> RESPONDED : Government delivers skrivelse
    PENDING --> OVERDUE : 4-month deadline elapsed, no response
    PENDING --> PARTIAL : Government responds partially
    OVERDUE --> RESPONDED : Belated skrivelse delivered
    PARTIAL --> RESPONDED : Full response delivered
    RESPONDED --> [*] : Follow-up complete

Library: scripts/rir-followups-client.ts

The rir-followups-client module provides all necessary tools:

import {
  calculateSkrivelseDeadline,  // Deadline from publish date (canonical spelling)
  deriveResponseStatus,       // Re-derive status vs today's date
  detectOverdueAlerts,        // Scan dataset for overdue records
  renderRirFollowUpTable,     // Render Markdown table
  injectRirTableIntoDocument, // Inject/replace table in a document
  filterByCommittee,          // Filter by Riksdag committee
  filterByStatus,             // Filter by response status
  filterByMinRiskLevel,       // Filter by minimum risk level
  validateRirRecord,          // Validate a single record
  validateRirDataset,         // Validate full dataset
  loadRirDataset,             // Load JSON from file (injectable I/O)
  saveRirDataset,             // Save JSON to file (injectable I/O)
} from './rir-followups-client.js';

CLI: scripts/fetch-rir-followups.ts

Run daily (or as part of the news workflow) to:

1. Fetch recent government skrivelser from data.riksdagen.se
2. Match against known RiR follow-up records
3. Update gov_response_status and response_skrivelse_id
4. Emit overdue alerts (exit code 1 with --alert flag)

# Check for overdue alerts (exit 1 if found)
npx tsx scripts/fetch-rir-followups.ts --alert

# Dry-run against a specific reference date
npx tsx scripts/fetch-rir-followups.ts --dry-run --date 2026-05-01

Injecting RiR Follow-Up Tables into Intelligence Assessments

Every intelligence-assessment document that cites a Riksrevisionen finding MUST include a follow-up status table. Use the injectRirTableIntoDocument helper or add the HTML markers manually:

<!-- RIR-FOLLOWUP-TABLE-START -->

## 🔍 Riksrevisionen Follow-Up Status

| RiR # | Title | Agency | Published | Skrivelse Deadline | Status | Days Overdue |
|-------|-------|--------|-----------|-------------------|--------|--------------|
| RiR 2026:6 | [Polisreform…](https://…) | Polismyndigheten | 2026-01-15 | 2026-05-15 | ⏳ PENDING | — |
| RiR 2025:18 | [Civilt försvar…](https://…) | MSB | 2025-12-10 | 2026-04-10 | 🚨 OVERDUE | ⚠️ 17 |

<!-- RIR-FOLLOWUP-TABLE-END -->

Riksdag API Search for Skrivelser

To fetch government responses manually via the riksdag-regering-mcp tools:

// Via search_dokument MCP tool
riksdag-regering-search_dokument({
  doktyp: 'skr',     // skrivelse (government communication)
  from_date: '2025-01-01',
  to_date: '2026-12-31',
  titel: 'Riksrevisionen'  // or specific RiR number
})

Workflow Integration

Wire the RiR follow-up tracker into intelligence-assessment generation by:

1. Daily fetch: Run fetch-rir-followups.ts before news generation to ensure data/rir-followups.json is up to date
2. Automatic injection: For any intelligence assessment document citing a RiR finding (detected via rir_report_id or rir_number keywords), call injectRirTableIntoDocument with the filtered relevant records
3. KJ propagation: When a record transitions to OVERDUE, update all active Key Judgements (KJs) that reference that RiR finding to reflect the accountability gap

Alert Thresholds

Days Overdue	Alert Level	Recommended Action
0	None	Normal monitoring
1–30	LOW	Note in intelligence assessment
31–90	MEDIUM	Create GitHub issue, escalate in weekly-review
91+	HIGH	Create GitHub issue immediately, flag in KJs

---

Hack23 ISMS Policy References

This skill implements requirements from:

• Secure Development Policy - Legislative data quality standards
• Information Security Policy - Data ethics in political monitoring
• Data Classification Policy - Political intelligence classification

References

Legislative Studies Literature

1. Cox, G. W., & McCubbins, M. D. (2005). Setting the Agenda: Responsible Party Government in the U.S. House of Representatives. Cambridge University Press.
2. Strøm, K. (1990). "A Behavioral Theory of Competitive Political Parties." American Journal of Political Science, 34(2), 565-598.
3. Döring, H. (1995). Parliaments and Majority Rule in Western Europe. Campus Verlag.

Database Intelligence Sources

• DATABASE_VIEW_INTELLIGENCE_CATALOG.md - Complete view documentation
• RISK_RULES_INTOP_OSINT.md - Risk rule specifications
• DATA_ANALYSIS_INTOP_OSINT.md - Analysis frameworks

---

🔗 Integration with agentic workflows & analysis artifacts

This skill is consumed by the 11 agentic news workflows in .github/workflows/news-*.md. The authoritative contract lives in .github/prompts/README.md; this skill supplies domain expertise on top of that contract.

• Analysis product → ai-driven-analysis-guide.md + every template in analysis/templates/.
• Required before any article: 9 core artifacts (14 for Tier-C) in analysis/daily/$ARTICLE_DATE/$SUBFOLDER/; 05-analysis-gate.md is the single blocking gate.
• gh-aw v0.69.3 docs: abridged · complete · blog series.

---

🔭 Horizon stratification

Authoritative source: .github/prompts/ext/long-horizon-forecasting.md. Runtime helper: scripts/horizon-context.ts.

Legislative monitoring uses horizon bands for bill-tracking and committee outlook confidence:

Band	Days	WEP language ceiling	Legislative application
72h	3	very likely / very unlikely	Scheduled plenary vote, confirmed committee agenda
week	7	likely / unlikely	Upcoming committee hearings, interpellation debates
month	30	likely / unlikely	Bill pipeline progress, budget amendment deadlines
quarter	90	roughly even / about even	Session-level legislative productivity outlook
year	365	roughly even; stronger requires ≥ 3 cycle-aged sources	Cross-session oversight effectiveness, GFS_COFOG spending trends
cycle	1460	roughly even / unlikely; never likely without ≥ 3 cycle-aged sources	Full mandate-period legislative agenda forecast
election	1460	scenario-driven; coalition outcomes never above "likely"	Post-election legislative programme scenarios

Cycle-rollover rules (±30 days of election anchor) are defined in .github/prompts/ext/cycle-rollover.md.