complexity-scorer

name	complexity-scorer
version	1.0.0
description	Score task complexity using keyword matching and heuristics
author	workspace-hub
category	bash
tags	["bash","complexity","scoring","keywords","analysis","classification"]
platforms	["linux","macos"]

Complexity Scorer

Patterns for scoring task complexity using keyword matching, context analysis, and configurable heuristics. Extracted from workspace-hub's model selection system.

When to Use This Skill

✅ Use when:

• Routing tasks to different handlers based on complexity
• Recommending resources (models, workers, time estimates)
• Prioritizing work items
• Auto-classifying incoming requests
• Building intelligent dispatchers

❌ Avoid when:

• Simple yes/no classification
• When ML-based classification is more appropriate
• Highly domain-specific scoring that requires expertise

Core Capabilities

1. Keyword-Based Scoring

Define keyword categories with weights:

#!/bin/bash
# ABOUTME: Keyword-based complexity scoring
# ABOUTME: Pattern from workspace-hub suggest_model.sh

# Keyword categories with associated complexity impact
# High complexity keywords (score +3)
HIGH_COMPLEXITY="architecture|refactor|design|security|complex|multi-file|algorithm|optimization|strategy|planning|cross-repository|performance|migration|integration"

# Medium complexity keywords (score +1)
MEDIUM_COMPLEXITY="implement|feature|bug|fix|code review|documentation|test|update|add|create|build|configure|setup"

# Low complexity keywords (score -2)
LOW_COMPLEXITY="check|status|simple|quick|template|list|grep|find|search|summary|validation|exists|show|display|count|verify"

# Score based on keywords
score_keywords() {
    local text="$1"
    local text_lower=$(echo "$text" | tr '[:upper:]' '[:lower:]')
    local score=0

    # Check high complexity first (mutually exclusive)
    if echo "$text_lower" | grep -qE "$HIGH_COMPLEXITY"; then
        ((score+=3))
    elif echo "$text_lower" | grep -qE "$MEDIUM_COMPLEXITY"; then
        ((score+=1))
    elif echo "" | grep -qE ; 
        ((score-=))
    

     
}


task=
score=$(score_keywords )
   

2. Multi-Factor Scoring

Combine multiple factors for better accuracy:

#!/bin/bash
# ABOUTME: Multi-factor complexity scoring
# ABOUTME: Combines keywords, length, context

# Factor weights
declare -A WEIGHTS=(
    ["keywords"]=3
    ["length"]=1
    ["urgency"]=1
    ["scope"]=2
)

# Score task length
score_length() {
    local text="$1"
    local word_count=$(echo "$text" | wc -w)

    if [[ $word_count -gt 20 ]]; then
        echo 2  # Long = more complex
    elif [[ $word_count -gt 10 ]]; then
        echo 1
    elif [[ $word_count -lt 5 ]]; then
        echo -1  # Short = simpler
    else
        echo 0
    fi
}

# Score urgency indicators
score_urgency() {
    local text="$1"
    local text_lower=$(echo "$text" | tr '[:upper:]' '[:lower:]')

    if echo  | grep -qE ; 
         2
       | grep -qE ; 
         1
    
         0
    
}


() {
     text=
     text_lower=$(  |   )

       | grep -qE ; 
         2
       | grep -qE ; 
         1
       | grep -qE ; 
         -1
    
         0
    
}


() {
     text=
     total=0

     keyword_score=$(score_keywords )
     length_score=$(score_length )
     urgency_score=$(score_urgency )
     scope_score=$(score_scope )

    total=$((keyword_score *  +
             length_score *  +
             urgency_score *  +
             scope_score * ))

     
}

3. Context-Aware Scoring

Adjust scores based on context (repository, domain, etc.):

#!/bin/bash
# ABOUTME: Context-aware complexity scoring
# ABOUTME: Adjusts based on repository tier, domain

# Repository tiers
TIER1_REPOS="workspace-hub|digitalmodel|energy|frontierdeepwater"
TIER2_REPOS="aceengineercode|assetutilities|worldenergydata|rock-oil-field"
TIER3_REPOS="doris|saipem|OGManufacturing|seanation"
PERSONAL_ACTIVE="aceengineer-admin|aceengineer-website"
PERSONAL_EXPERIMENTAL="hobbies|sd-work|acma-projects"

# Get repository tier
get_repo_tier() {
    local repo="$1"

    if echo "$repo" | grep -qE "$TIER1_REPOS"; then
        echo "tier1"
    elif echo "$repo" | grep -qE "$TIER2_REPOS"; then
        echo "tier2"
    elif echo "$repo" | grep -qE "$TIER3_REPOS"; then
        echo "tier3"
    elif echo "$repo" | grep -qE "$PERSONAL_ACTIVE"; then
        echo 
       | grep -qE ; 
         
    
         
    
}


() {
     base_score=
     repo=
     adjusted=

     tier=$(get_repo_tier )

      
        tier1)
            
            ((adjusted+=))
            ;;
        tier3|personal_experimental)
            
            ((adjusted-=))
            ;;
    

     
}


() {
     task=
     repo=

     base_score=$(calculate_complexity )
     final_score=$(adjust_for_context  )

     
}

4. Score Classification

Map scores to categories:

#!/bin/bash
# ABOUTME: Map complexity scores to categories
# ABOUTME: Provides actionable classifications

# Classification thresholds
THRESHOLD_HIGH=5
THRESHOLD_MEDIUM=1
THRESHOLD_LOW=-2

# Classify score
classify_complexity() {
    local score="$1"

    if [[ $score -ge $THRESHOLD_HIGH ]]; then
        echo "high"
    elif [[ $score -ge $THRESHOLD_MEDIUM ]]; then
        echo "medium"
    elif [[ $score -ge $THRESHOLD_LOW ]]; then
        echo "low"
    else
        echo "trivial"
    fi
}

# Get recommendation based on classification
get_recommendation() {
    local score="$1"
    local classification=$(classify_complexity "$score")

    case "$classification" in
        high)
            echo "OPUS"
            echo "Use for complex architecture, multi-file refactoring, security analysis"
            ;;
        medium)
             
             
            ;;
        low)
             
             
            ;;
        trivial)
             
             
            ;;
    
}


() {
     score=
     classification=$(classify_complexity )

      
        high)      ;;
        medium)    ;;
        low)       ;;
        trivial)   ;;
    
}

5. Confidence Scoring

Add confidence to scoring:

#!/bin/bash
# ABOUTME: Confidence scoring for complexity estimates
# ABOUTME: Indicates reliability of the score

# Calculate confidence
calculate_confidence() {
    local text="$1"
    local confidence=50  # Base confidence

    local word_count=$(echo "$text" | wc -w)
    local keyword_matches=0
    local text_lower=$(echo "$text" | tr '[:upper:]' '[:lower:]')

    # More words = higher confidence (more context)
    if [[ $word_count -gt 15 ]]; then
        ((confidence+=20))
    elif [[ $word_count -gt 8 ]]; then
        ((confidence+=10))
    elif [[ $word_count -lt 3 ]]; then
        ((confidence-=20))
    fi

    # Keyword matches boost confidence
    for pattern in "$HIGH_COMPLEXITY" "$MEDIUM_COMPLEXITY" "$LOW_COMPLEXITY"; do
        if echo "" | grep -qE ; 
            ((keyword_matches++))
        
    

    ((confidence += keyword_matches * ))

    
    [[  -gt 100 ]] && confidence=100
    [[  -lt 10 ]] && confidence=10

     
}


() {
     confidence=

     [[  -ge 80 ]]; 
         
     [[  -ge 60 ]]; 
         
     [[  -ge 40 ]]; 
         
    
         
    
}

6. Historical Learning

Adjust based on past accuracy:

#!/bin/bash
# ABOUTME: Historical learning for complexity scoring
# ABOUTME: Track and adjust based on actual outcomes

HISTORY_FILE="${HOME}/.complexity-scorer/history.log"

# Log prediction vs actual
log_outcome() {
    local task="$1"
    local predicted_score="$2"
    local actual_complexity="$3"  # user-provided feedback
    local timestamp=$(date '+%Y-%m-%d_%H:%M:%S')

    mkdir -p "$(dirname "$HISTORY_FILE")"
    echo "${timestamp}|${predicted_score}|${actual_complexity}|${task}" >> "$HISTORY_FILE"
}

# Calculate prediction accuracy
calculate_accuracy() {
    local correct=0
    local total=0

    while IFS='|' read -r ts predicted actual task; do
        [[ "$ts" =~ ^#.*$ ]] && continue
        [[ -z "$ts" ]] && continue

        ((total++))

        local predicted_class=$(classify_complexity )
         [[  ==  ]]; 
            ((correct++))
        
     < 

     [[  -gt 0 ]]; 
         $((correct *  / total))
    
         0
    
}


() {
     

     IFS=  -r ts predicted actual task; 
         predicted_class=$(classify_complexity )
         [[  !=  ]]; 
             
             
             
        
     < 
}

Complete Example: Task Complexity Scorer

#!/bin/bash
# ABOUTME: Complete task complexity scoring system
# ABOUTME: Multi-factor scoring with recommendations

set -e

# ─────────────────────────────────────────────────────────────────
# Configuration
# ─────────────────────────────────────────────────────────────────

# Colors
GREEN='\033[0;32m'
BLUE='\033[0;34m'
YELLOW='\033[1;33m'
CYAN='\033[0;36m'
RED='\033[0;31m'
NC='\033[0m'

# Keyword patterns
OPUS_KEYWORDS="architecture|refactor|design|security|complex|multi-file|algorithm|optimization|strategy|planning|cross-repository|performance|migration"
SONNET_KEYWORDS="implement|feature|bug|fix|code review|documentation|test|update|add|create|build"
HAIKU_KEYWORDS="check|status|simple|quick|template|list|grep|find|search|summary|validation|exists|show|display"

# Repository tiers
WORK_TIER1="workspace-hub|digitalmodel|energy|frontierdeepwater"
WORK_TIER2="aceengineercode|assetutilities|worldenergydata"
WORK_TIER3="doris|saipem|OGManufacturing"
PERSONAL="hobbies|sd-work|acma-projects"

# ─────────────────────────────────────────────────────────────────
# Scoring Functions
# ─────────────────────────────────────────────────────────────────

score_task() {
    local task="$1"
    local repo="${2:-}"
    local task_lower=$(echo "$task" | tr '[:upper:]' )
     score=0

    
       | grep -qE ; 
        ((score+=))
       | grep -qE ; 
        ((score+=))
       | grep -qE ; 
        ((score-=))
    

    
     word_count=$(  |  -w)
     [[  -gt 15 ]]; 
        ((score+=))
     [[  -lt 5 ]]; 
        ((score-=))
    

    
     [[ -n  ]]; 
           | grep -qE ; 
            ((score+=))
           | grep -qE ; 
            ((score-=))
        
    

     
}

() {
     score=

     [[  -ge 3 ]]; 
         
     [[  -ge 0 ]]; 
         
    
         
    
}

() {
     repo=

       | grep -qE ; 
         
       | grep -qE ; 
         
       | grep -qE ; 
         
       | grep -qE ; 
         
    
         
    
}





() {
     task=
     repo=
     score=$(score_task  )
     recommendation=$(get_recommendation )
     tier=$(get_tier_name )

    
     color
      
        OPUS)   color= ;;
        SONNET) color= ;;
        HAIKU)  color= ;;
    

     
     -e 
     -e 
     -e 
     

    [[ -n  ]] &&  -e 
     -e 
     -e 
     
     -e 
     

    
     -e 
     task_lower=$(  |   )

       | grep -qE ; 
         
       | grep -qE ; 
         
       | grep -qE ; 
         
    

    [[ -n  ]] &&  

     
     -e 
}





() {
     << 
}

REPO=
SCORE_ONLY=
JSON_OUTPUT=

 [[  -gt 0 ]]; 
      
        -r|--repo)
            REPO=
             2
            ;;
        -s|--score-only)
            SCORE_ONLY=
            
            ;;
        -j|--json)
            JSON_OUTPUT=
            
            ;;
        -h|--)
            show_usage
             0
            ;;
        -*)
             
            show_usage
             1
            ;;
        *)
            
            ;;
    


TASK=

 [[ -z  ]]; 
     -p  TASK


 [[ -z  ]]; 
     
     1


 [[  ==  ]]; 
    score_task  
 [[  ==  ]]; 
    score=$(score_task  )
    rec=$(get_recommendation )
     << 

    display_result  

Best Practices

1. Tune Keywords - Adjust patterns based on your domain
2. Use Multiple Factors - Don't rely on keywords alone
3. Add Confidence - Help users understand reliability
4. Learn from Feedback - Track accuracy and improve
5. Context Matters - Same task may have different complexity in different contexts

Resources

• Regular Expressions in Bash
• Pattern Matching

---

Version History

• 1.0.0 (2026-01-14): Initial release - extracted from workspace-hub suggest_model.sh