aiox-core 5.0.3 → 5.0.4

package/pro/squads/squad-creator-pro/scripts/assess-sources.sh

@@ -1,443 +1,443 @@
-#!/bin/bash
-# assess-sources.sh - Deterministic source quality assessment (no LLM needed)
-# Usage: ./assess-sources.sh <sources-dir> [--json] [--verbose]
-#
-# Loops through source files, applies 25 binary checkpoints per source
-# across 5 dimensions (Autenticidade, Profundidade, Atualidade, Unicidade, Completude)
-
-set -e
-
-SOURCES_DIR=""
-OUTPUT_FORMAT="yaml"
-VERBOSE=false
-
-# Parse arguments
-while [[ $# -gt 0 ]]; do
-    case $1 in
-        --json)
-            OUTPUT_FORMAT="json"
-            shift
-            ;;
-        --verbose)
-            VERBOSE=true
-            shift
-            ;;
-        -*)
-            echo "Unknown option: $1"
-            exit 1
-            ;;
-        *)
-            SOURCES_DIR="$1"
-            shift
-            ;;
-    esac
-done
-
-if [[ -z "$SOURCES_DIR" ]]; then
-    echo "Usage: $0 <sources-dir> [--json] [--verbose]"
-    echo ""
-    echo "Assesses quality of source files across 5 dimensions (25 checkpoints)."
-    echo "Classifies each source as ouro (gold), mixed, or bronze."
-    echo ""
-    echo "Options:"
-    echo "  --json      Output JSON instead of YAML"
-    echo "  --verbose   Show per-checkpoint details for each source"
-    exit 1
-fi
-
-if [[ ! -d "$SOURCES_DIR" ]]; then
-    echo "Error: Directory not found: $SOURCES_DIR"
-    exit 1
-fi
-
-DATE=$(date -Iseconds)
-
-# Find all source files
-mapfile -t SOURCE_FILES < <(find "$SOURCES_DIR" -type f \( -name "*.md" -o -name "*.yaml" -o -name "*.yml" -o -name "*.txt" \) | sort)
-
-TOTAL_FILES=${#SOURCE_FILES[@]}
-
-if [[ $TOTAL_FILES -eq 0 ]]; then
-    echo "Error: No source files (.md, .yaml, .yml, .txt) found in $SOURCES_DIR"
-    exit 1
-fi
-
-# Counters
-OURO_COUNT=0
-MIXED_COUNT=0
-BRONZE_COUNT=0
-CROWN_COUNT=0
-ALL_SCORES=()
-ALL_TIERS=()
-ALL_NAMES=()
-
-# Per-source YAML/JSON accumulator
-SOURCE_REPORTS=""
-
-for FILE in "${SOURCE_FILES[@]}"; do
-    FNAME=$(basename "$FILE")
-    LINE_COUNT=$(wc -l < "$FILE" | tr -d ' ')
-
-    # ========================================================================
-    # D1: Autenticidade (5 checks)
-    # ========================================================================
-    D1=0
-
-    # D1.1: Spontaneity markers ("honestly", "I was wrong", "truth is")
-    if grep -qiE "(honestly|I was wrong|truth is|to be honest|frankly|confession)" "$FILE"; then
-        D1=$((D1 + 1))
-        D1_1="PASS"
-    else
-        D1_1="FAIL"
-    fi
-
-    # D1.2: Vulnerability markers ("struggle", "failure", "mistake")
-    if grep -qiE "(struggle|failure|mistake|I failed|I lost|painful|embarrass)" "$FILE"; then
-        D1=$((D1 + 1))
-        D1_2="PASS"
-    else
-        D1_2="FAIL"
-    fi
-
-    # D1.3: Difficult questions acknowledged
-    if grep -qiE "(difficult question|hard question|I don.t know|uncertain|not sure|gray area)" "$FILE"; then
-        D1=$((D1 + 1))
-        D1_3="PASS"
-    else
-        D1_3="FAIL"
-    fi
-
-    # D1.4: Nuanced answers (not black/white)
-    if grep -qiE "(depends on|it depends|nuance|on the other hand|however|but also|both true)" "$FILE"; then
-        D1=$((D1 + 1))
-        D1_4="PASS"
-    else
-        D1_4="FAIL"
-    fi
-
-    # D1.5: Personal storytelling
-    if grep -qiE "(I remember|when I was|my experience|I learned|my story|years ago)" "$FILE"; then
-        D1=$((D1 + 1))
-        D1_5="PASS"
-    else
-        D1_5="FAIL"
-    fi
-
-    # ========================================================================
-    # D2: Profundidade (5 checks)
-    # ========================================================================
-    D2=0
-
-    # D2.1: Explains why (not just what)
-    if grep -qiE "(because|the reason|why this matters|this is important|the key insight)" "$FILE"; then
-        D2=$((D2 + 1))
-        D2_1="PASS"
-    else
-        D2_1="FAIL"
-    fi
-
-    # D2.2: Framework present
-    if grep -qiE "(framework|model|system|methodology|approach|process|step [0-9])" "$FILE"; then
-        D2=$((D2 + 1))
-        D2_2="PASS"
-    else
-        D2_2="FAIL"
-    fi
-
-    # D2.3: Specific examples with numbers
-    if grep -qE "[0-9]+%" "$FILE" || grep -qiE "(\$[0-9]|[0-9]+ (people|clients|companies|students|years))" "$FILE"; then
-        D2=$((D2 + 1))
-        D2_3="PASS"
-    else
-        D2_3="FAIL"
-    fi
-
-    # D2.4: Trade-offs discussed
-    if grep -qiE "(trade.off|downside|cost of|sacrifice|risk|limitation|the problem with)" "$FILE"; then
-        D2=$((D2 + 1))
-        D2_4="PASS"
-    else
-        D2_4="FAIL"
-    fi
-
-    # D2.5: Cross-domain connections
-    if grep -qiE "(like in|similar to|just as|reminds me of|analogy|compare|connection)" "$FILE"; then
-        D2=$((D2 + 1))
-        D2_5="PASS"
-    else
-        D2_5="FAIL"
-    fi
-
-    # ========================================================================
-    # D3: Atualidade (5 checks)
-    # ========================================================================
-    D3=0
-
-    # D3.1: Date >= 2023 (or recent context)
-    if grep -qE "(202[3-9]|203[0-9])" "$FILE"; then
-        D3=$((D3 + 1))
-        D3_1="PASS"
-    else
-        D3_1="FAIL"
-    fi
-
-    # D3.2: Current context references (AI, GPT, remote work, etc.)
-    if grep -qiE "(AI|artificial intelligence|GPT|ChatGPT|remote|hybrid work|creator economy|no.code)" "$FILE"; then
-        D3=$((D3 + 1))
-        D3_2="PASS"
-    else
-        D3_2="FAIL"
-    fi
-
-    # D3.3: Not contradicted (no "outdated" or "deprecated" markers)
-    if ! grep -qiE "(outdated|deprecated|no longer|obsolete|was true but)" "$FILE"; then
-        D3=$((D3 + 1))
-        D3_3="PASS"
-    else
-        D3_3="FAIL"
-    fi
-
-    # D3.4: Shows evolution (mentions change, growth, update)
-    if grep -qiE "(evolved|changed|updated|used to|now I|shifted|pivot)" "$FILE"; then
-        D3=$((D3 + 1))
-        D3_4="PASS"
-    else
-        D3_4="FAIL"
-    fi
-
-    # D3.5: Applicable now (actionable content)
-    if grep -qiE "(you can|try this|start with|actionable|implement|apply|do this)" "$FILE"; then
-        D3=$((D3 + 1))
-        D3_5="PASS"
-    else
-        D3_5="FAIL"
-    fi
-
-    # ========================================================================
-    # D4: Unicidade (5 checks)
-    # ========================================================================
-    D4=0
-
-    # D4.1: Contra-mainstream position
-    if grep -qiE "(most people|everyone says|conventional|contrarian|unpopular|disagree|wrong about)" "$FILE"; then
-        D4=$((D4 + 1))
-        D4_1="PASS"
-    else
-        D4_1="FAIL"
-    fi
-
-    # D4.2: Unique vocabulary (coined terms)
-    if grep -qiE "(I call it|what I call|my term|I coined|I define)" "$FILE"; then
-        D4=$((D4 + 1))
-        D4_2="PASS"
-    else
-        D4_2="FAIL"
-    fi
-
-    # D4.3: Original framework name
-    # Check for capitalized multi-word names that look like named frameworks
-    if grep -qE "[A-Z][a-z]+\s+[A-Z][a-z]+" "$FILE" && grep -qiE "(framework|method|system|model|approach)" "$FILE"; then
-        D4=$((D4 + 1))
-        D4_3="PASS"
-    else
-        D4_3="FAIL"
-    fi
-
-    # D4.4: Counter-intuitive insight
-    if grep -qiE "(counter.intuitive|surprising|paradox|opposite|actually|seems wrong but)" "$FILE"; then
-        D4=$((D4 + 1))
-        D4_4="PASS"
-    else
-        D4_4="FAIL"
-    fi
-
-    # D4.5: Singular perspective
-    if grep -qiE "(my perspective|in my view|I believe|my philosophy|my approach|the way I see)" "$FILE"; then
-        D4=$((D4 + 1))
-        D4_5="PASS"
-    else
-        D4_5="FAIL"
-    fi
-
-    # ========================================================================
-    # D5: Completude (5 checks)
-    # ========================================================================
-    D5=0
-
-    # D5.1: Has step-by-step
-    if grep -qiE "(step [0-9]|step-by-step|first.*then.*finally|1\.|2\.|3\.)" "$FILE"; then
-        D5=$((D5 + 1))
-        D5_1="PASS"
-    else
-        D5_1="FAIL"
-    fi
-
-    # D5.2: Has decision rules
-    if grep -qiE "(if.*then|when.*do|rule:|decision:|criteria:|threshold)" "$FILE"; then
-        D5=$((D5 + 1))
-        D5_2="PASS"
-    else
-        D5_2="FAIL"
-    fi
-
-    # D5.3: Has real examples
-    if grep -qiE "(for example|example:|e\.g\.|such as|case study|instance)" "$FILE"; then
-        D5=$((D5 + 1))
-        D5_3="PASS"
-    else
-        D5_3="FAIL"
-    fi
-
-    # D5.4: Discusses edge cases
-    if grep -qiE "(edge case|exception|unless|except when|caveat|warning|watch out)" "$FILE"; then
-        D5=$((D5 + 1))
-        D5_4="PASS"
-    else
-        D5_4="FAIL"
-    fi
-
-    # D5.5: Adequate length (>200 lines or indication of substantial content)
-    if [[ "$LINE_COUNT" -gt 200 ]]; then
-        D5=$((D5 + 1))
-        D5_5="PASS"
-    else
-        # Check word count as fallback
-        WORD_COUNT=$(wc -w < "$FILE" | tr -d ' ')
-        if [[ "$WORD_COUNT" -gt 1500 ]]; then
-            D5=$((D5 + 1))
-            D5_5="PASS"
-        else
-            D5_5="FAIL"
-        fi
-    fi
-
-    # ========================================================================
-    # SCORING
-    # ========================================================================
-    TOTAL_CHECKS=$((D1 + D2 + D3 + D4 + D5))
-    # Average across 5 dimensions (each 0-5, average 0-5)
-    AVG_SCORE=$(echo "scale=1; ($D1 + $D2 + $D3 + $D4 + $D5) / 5" | bc)
-
-    # Tier classification
-    # Special rule: D1 < 3 → force bronze
-    if [[ "$D1" -lt 3 ]]; then
-        TIER="bronze"
-        TIER_REASON="D1 Autenticidade < 3 (forced bronze)"
-    elif [[ "$D1" -ge 4 && "$D2" -ge 4 && "$D3" -ge 4 && "$D4" -ge 4 && "$D5" -ge 4 ]]; then
-        TIER="crown_jewel"
-        TIER_REASON="All dimensions >= 4"
-        CROWN_COUNT=$((CROWN_COUNT + 1))
-        OURO_COUNT=$((OURO_COUNT + 1))
-    elif (( $(echo "$AVG_SCORE >= 4.0" | bc -l) )); then
-        TIER="ouro"
-        TIER_REASON="Average >= 4.0"
-        OURO_COUNT=$((OURO_COUNT + 1))
-    elif (( $(echo "$AVG_SCORE >= 3.0" | bc -l) )); then
-        TIER="mixed"
-        TIER_REASON="Average 3.0-3.9"
-        MIXED_COUNT=$((MIXED_COUNT + 1))
-    else
-        TIER="bronze"
-        TIER_REASON="Average < 3.0"
-        BRONZE_COUNT=$((BRONZE_COUNT + 1))
-    fi
-
-    ALL_SCORES+=("$AVG_SCORE")
-    ALL_TIERS+=("$TIER")
-    ALL_NAMES+=("$FNAME")
-
-    # Accumulate per-source report
-    if [[ "$OUTPUT_FORMAT" == "yaml" ]]; then
-        SOURCE_REPORTS+="    - file: \"$FNAME\""$'\n'
-        SOURCE_REPORTS+="      lines: $LINE_COUNT"$'\n'
-        SOURCE_REPORTS+="      tier: \"$TIER\""$'\n'
-        SOURCE_REPORTS+="      tier_reason: \"$TIER_REASON\""$'\n'
-        SOURCE_REPORTS+="      average_score: $AVG_SCORE"$'\n'
-        SOURCE_REPORTS+="      dimensions:"$'\n'
-        SOURCE_REPORTS+="        autenticidade: { score: $D1, checks: [$D1_1, $D1_2, $D1_3, $D1_4, $D1_5] }"$'\n'
-        SOURCE_REPORTS+="        profundidade: { score: $D2, checks: [$D2_1, $D2_2, $D2_3, $D2_4, $D2_5] }"$'\n'
-        SOURCE_REPORTS+="        atualidade: { score: $D3, checks: [$D3_1, $D3_2, $D3_3, $D3_4, $D3_5] }"$'\n'
-        SOURCE_REPORTS+="        unicidade: { score: $D4, checks: [$D4_1, $D4_2, $D4_3, $D4_4, $D4_5] }"$'\n'
-        SOURCE_REPORTS+="        completude: { score: $D5, checks: [$D5_1, $D5_2, $D5_3, $D5_4, $D5_5] }"$'\n'
-    fi
-
-    if [[ "$VERBOSE" == true ]]; then
-        echo "  [VERBOSE] $FNAME: D1=$D1 D2=$D2 D3=$D3 D4=$D4 D5=$D5 avg=$AVG_SCORE tier=$TIER" >&2
-    fi
-done
-
-# ============================================================================
-# AGGREGATION
-# ============================================================================
-if [[ $TOTAL_FILES -gt 0 ]]; then
-    OURO_PCT=$((OURO_COUNT * 100 / TOTAL_FILES))
-else
-    OURO_PCT=0
-fi
-
-# Overall quality
-if [[ "$OURO_PCT" -ge 80 ]]; then
-    OVERALL="excelente"
-elif [[ "$OURO_PCT" -ge 60 ]]; then
-    OVERALL="aceitavel"
-elif [[ "$OURO_PCT" -ge 40 ]]; then
-    OVERALL="parcial"
-else
-    OVERALL="critico"
-fi
-
-# ============================================================================
-# OUTPUT
-# ============================================================================
-if [[ "$OUTPUT_FORMAT" == "json" ]]; then
-    echo "{"
-    echo "  \"source_assessment\": {"
-    echo "    \"directory\": \"$SOURCES_DIR\","
-    echo "    \"date\": \"$DATE\","
-    echo "    \"summary\": {"
-    echo "      \"total_sources\": $TOTAL_FILES,"
-    echo "      \"ouro_count\": $OURO_COUNT,"
-    echo "      \"mixed_count\": $MIXED_COUNT,"
-    echo "      \"bronze_count\": $BRONZE_COUNT,"
-    echo "      \"crown_jewels\": $CROWN_COUNT,"
-    echo "      \"ouro_percentage\": \"${OURO_PCT}%\","
-    echo "      \"overall_quality\": \"$OVERALL\""
-    echo "    },"
-    echo "    \"sources\": ["
-
-    for i in "${!ALL_NAMES[@]}"; do
-        COMMA=""
-        if [[ $i -lt $((${#ALL_NAMES[@]} - 1)) ]]; then
-            COMMA=","
-        fi
-        echo "      {\"file\": \"${ALL_NAMES[$i]}\", \"tier\": \"${ALL_TIERS[$i]}\", \"score\": ${ALL_SCORES[$i]}}$COMMA"
-    done
-
-    echo "    ]"
-    echo "  }"
-    echo "}"
-else
-    # YAML output
-    echo "# Source Assessment Report"
-    echo "# Generated: $DATE"
-    echo "# Directory: $SOURCES_DIR"
-    echo ""
-    echo "source_assessment:"
-    echo "  directory: \"$SOURCES_DIR\""
-    echo "  date: \"$DATE\""
-    echo ""
-    echo "  summary:"
-    echo "    total_sources: $TOTAL_FILES"
-    echo "    ouro_count: $OURO_COUNT"
-    echo "    mixed_count: $MIXED_COUNT"
-    echo "    bronze_count: $BRONZE_COUNT"
-    echo "    crown_jewels: $CROWN_COUNT"
-    echo "    ouro_percentage: \"${OURO_PCT}%\""
-    echo "    overall_quality: \"$OVERALL\""
-    echo ""
-    echo "  sources:"
-    echo -n "$SOURCE_REPORTS"
-    echo ""
-    echo "# Cost: \$0 (zero tokens)"
-    echo "# Time: <${TOTAL_FILES} seconds"
-fi
+#!/bin/bash
+# assess-sources.sh - Deterministic source quality assessment (no LLM needed)
+# Usage: ./assess-sources.sh <sources-dir> [--json] [--verbose]
+#
+# Loops through source files, applies 25 binary checkpoints per source
+# across 5 dimensions (Autenticidade, Profundidade, Atualidade, Unicidade, Completude)
+
+set -e
+
+SOURCES_DIR=""
+OUTPUT_FORMAT="yaml"
+VERBOSE=false
+
+# Parse arguments
+while [[ $# -gt 0 ]]; do
+    case $1 in
+        --json)
+            OUTPUT_FORMAT="json"
+            shift
+            ;;
+        --verbose)
+            VERBOSE=true
+            shift
+            ;;
+        -*)
+            echo "Unknown option: $1"
+            exit 1
+            ;;
+        *)
+            SOURCES_DIR="$1"
+            shift
+            ;;
+    esac
+done
+
+if [[ -z "$SOURCES_DIR" ]]; then
+    echo "Usage: $0 <sources-dir> [--json] [--verbose]"
+    echo ""
+    echo "Assesses quality of source files across 5 dimensions (25 checkpoints)."
+    echo "Classifies each source as ouro (gold), mixed, or bronze."
+    echo ""
+    echo "Options:"
+    echo "  --json      Output JSON instead of YAML"
+    echo "  --verbose   Show per-checkpoint details for each source"
+    exit 1
+fi
+
+if [[ ! -d "$SOURCES_DIR" ]]; then
+    echo "Error: Directory not found: $SOURCES_DIR"
+    exit 1
+fi
+
+DATE=$(date -Iseconds)
+
+# Find all source files
+mapfile -t SOURCE_FILES < <(find "$SOURCES_DIR" -type f \( -name "*.md" -o -name "*.yaml" -o -name "*.yml" -o -name "*.txt" \) | sort)
+
+TOTAL_FILES=${#SOURCE_FILES[@]}
+
+if [[ $TOTAL_FILES -eq 0 ]]; then
+    echo "Error: No source files (.md, .yaml, .yml, .txt) found in $SOURCES_DIR"
+    exit 1
+fi
+
+# Counters
+OURO_COUNT=0
+MIXED_COUNT=0
+BRONZE_COUNT=0
+CROWN_COUNT=0
+ALL_SCORES=()
+ALL_TIERS=()
+ALL_NAMES=()
+
+# Per-source YAML/JSON accumulator
+SOURCE_REPORTS=""
+
+for FILE in "${SOURCE_FILES[@]}"; do
+    FNAME=$(basename "$FILE")
+    LINE_COUNT=$(wc -l < "$FILE" | tr -d ' ')
+
+    # ========================================================================
+    # D1: Autenticidade (5 checks)
+    # ========================================================================
+    D1=0
+
+    # D1.1: Spontaneity markers ("honestly", "I was wrong", "truth is")
+    if grep -qiE "(honestly|I was wrong|truth is|to be honest|frankly|confession)" "$FILE"; then
+        D1=$((D1 + 1))
+        D1_1="PASS"
+    else
+        D1_1="FAIL"
+    fi
+
+    # D1.2: Vulnerability markers ("struggle", "failure", "mistake")
+    if grep -qiE "(struggle|failure|mistake|I failed|I lost|painful|embarrass)" "$FILE"; then
+        D1=$((D1 + 1))
+        D1_2="PASS"
+    else
+        D1_2="FAIL"
+    fi
+
+    # D1.3: Difficult questions acknowledged
+    if grep -qiE "(difficult question|hard question|I don.t know|uncertain|not sure|gray area)" "$FILE"; then
+        D1=$((D1 + 1))
+        D1_3="PASS"
+    else
+        D1_3="FAIL"
+    fi
+
+    # D1.4: Nuanced answers (not black/white)
+    if grep -qiE "(depends on|it depends|nuance|on the other hand|however|but also|both true)" "$FILE"; then
+        D1=$((D1 + 1))
+        D1_4="PASS"
+    else
+        D1_4="FAIL"
+    fi
+
+    # D1.5: Personal storytelling
+    if grep -qiE "(I remember|when I was|my experience|I learned|my story|years ago)" "$FILE"; then
+        D1=$((D1 + 1))
+        D1_5="PASS"
+    else
+        D1_5="FAIL"
+    fi
+
+    # ========================================================================
+    # D2: Profundidade (5 checks)
+    # ========================================================================
+    D2=0
+
+    # D2.1: Explains why (not just what)
+    if grep -qiE "(because|the reason|why this matters|this is important|the key insight)" "$FILE"; then
+        D2=$((D2 + 1))
+        D2_1="PASS"
+    else
+        D2_1="FAIL"
+    fi
+
+    # D2.2: Framework present
+    if grep -qiE "(framework|model|system|methodology|approach|process|step [0-9])" "$FILE"; then
+        D2=$((D2 + 1))
+        D2_2="PASS"
+    else
+        D2_2="FAIL"
+    fi
+
+    # D2.3: Specific examples with numbers
+    if grep -qE "[0-9]+%" "$FILE" || grep -qiE "(\$[0-9]|[0-9]+ (people|clients|companies|students|years))" "$FILE"; then
+        D2=$((D2 + 1))
+        D2_3="PASS"
+    else
+        D2_3="FAIL"
+    fi
+
+    # D2.4: Trade-offs discussed
+    if grep -qiE "(trade.off|downside|cost of|sacrifice|risk|limitation|the problem with)" "$FILE"; then
+        D2=$((D2 + 1))
+        D2_4="PASS"
+    else
+        D2_4="FAIL"
+    fi
+
+    # D2.5: Cross-domain connections
+    if grep -qiE "(like in|similar to|just as|reminds me of|analogy|compare|connection)" "$FILE"; then
+        D2=$((D2 + 1))
+        D2_5="PASS"
+    else
+        D2_5="FAIL"
+    fi
+
+    # ========================================================================
+    # D3: Atualidade (5 checks)
+    # ========================================================================
+    D3=0
+
+    # D3.1: Date >= 2023 (or recent context)
+    if grep -qE "(202[3-9]|203[0-9])" "$FILE"; then
+        D3=$((D3 + 1))
+        D3_1="PASS"
+    else
+        D3_1="FAIL"
+    fi
+
+    # D3.2: Current context references (AI, GPT, remote work, etc.)
+    if grep -qiE "(AI|artificial intelligence|GPT|ChatGPT|remote|hybrid work|creator economy|no.code)" "$FILE"; then
+        D3=$((D3 + 1))
+        D3_2="PASS"
+    else
+        D3_2="FAIL"
+    fi
+
+    # D3.3: Not contradicted (no "outdated" or "deprecated" markers)
+    if ! grep -qiE "(outdated|deprecated|no longer|obsolete|was true but)" "$FILE"; then
+        D3=$((D3 + 1))
+        D3_3="PASS"
+    else
+        D3_3="FAIL"
+    fi
+
+    # D3.4: Shows evolution (mentions change, growth, update)
+    if grep -qiE "(evolved|changed|updated|used to|now I|shifted|pivot)" "$FILE"; then
+        D3=$((D3 + 1))
+        D3_4="PASS"
+    else
+        D3_4="FAIL"
+    fi
+
+    # D3.5: Applicable now (actionable content)
+    if grep -qiE "(you can|try this|start with|actionable|implement|apply|do this)" "$FILE"; then
+        D3=$((D3 + 1))
+        D3_5="PASS"
+    else
+        D3_5="FAIL"
+    fi
+
+    # ========================================================================
+    # D4: Unicidade (5 checks)
+    # ========================================================================
+    D4=0
+
+    # D4.1: Contra-mainstream position
+    if grep -qiE "(most people|everyone says|conventional|contrarian|unpopular|disagree|wrong about)" "$FILE"; then
+        D4=$((D4 + 1))
+        D4_1="PASS"
+    else
+        D4_1="FAIL"
+    fi
+
+    # D4.2: Unique vocabulary (coined terms)
+    if grep -qiE "(I call it|what I call|my term|I coined|I define)" "$FILE"; then
+        D4=$((D4 + 1))
+        D4_2="PASS"
+    else
+        D4_2="FAIL"
+    fi
+
+    # D4.3: Original framework name
+    # Check for capitalized multi-word names that look like named frameworks
+    if grep -qE "[A-Z][a-z]+\s+[A-Z][a-z]+" "$FILE" && grep -qiE "(framework|method|system|model|approach)" "$FILE"; then
+        D4=$((D4 + 1))
+        D4_3="PASS"
+    else
+        D4_3="FAIL"
+    fi
+
+    # D4.4: Counter-intuitive insight
+    if grep -qiE "(counter.intuitive|surprising|paradox|opposite|actually|seems wrong but)" "$FILE"; then
+        D4=$((D4 + 1))
+        D4_4="PASS"
+    else
+        D4_4="FAIL"
+    fi
+
+    # D4.5: Singular perspective
+    if grep -qiE "(my perspective|in my view|I believe|my philosophy|my approach|the way I see)" "$FILE"; then
+        D4=$((D4 + 1))
+        D4_5="PASS"
+    else
+        D4_5="FAIL"
+    fi
+
+    # ========================================================================
+    # D5: Completude (5 checks)
+    # ========================================================================
+    D5=0
+
+    # D5.1: Has step-by-step
+    if grep -qiE "(step [0-9]|step-by-step|first.*then.*finally|1\.|2\.|3\.)" "$FILE"; then
+        D5=$((D5 + 1))
+        D5_1="PASS"
+    else
+        D5_1="FAIL"
+    fi
+
+    # D5.2: Has decision rules
+    if grep -qiE "(if.*then|when.*do|rule:|decision:|criteria:|threshold)" "$FILE"; then
+        D5=$((D5 + 1))
+        D5_2="PASS"
+    else
+        D5_2="FAIL"
+    fi
+
+    # D5.3: Has real examples
+    if grep -qiE "(for example|example:|e\.g\.|such as|case study|instance)" "$FILE"; then
+        D5=$((D5 + 1))
+        D5_3="PASS"
+    else
+        D5_3="FAIL"
+    fi
+
+    # D5.4: Discusses edge cases
+    if grep -qiE "(edge case|exception|unless|except when|caveat|warning|watch out)" "$FILE"; then
+        D5=$((D5 + 1))
+        D5_4="PASS"
+    else
+        D5_4="FAIL"
+    fi
+
+    # D5.5: Adequate length (>200 lines or indication of substantial content)
+    if [[ "$LINE_COUNT" -gt 200 ]]; then
+        D5=$((D5 + 1))
+        D5_5="PASS"
+    else
+        # Check word count as fallback
+        WORD_COUNT=$(wc -w < "$FILE" | tr -d ' ')
+        if [[ "$WORD_COUNT" -gt 1500 ]]; then
+            D5=$((D5 + 1))
+            D5_5="PASS"
+        else
+            D5_5="FAIL"
+        fi
+    fi
+
+    # ========================================================================
+    # SCORING
+    # ========================================================================
+    TOTAL_CHECKS=$((D1 + D2 + D3 + D4 + D5))
+    # Average across 5 dimensions (each 0-5, average 0-5)
+    AVG_SCORE=$(echo "scale=1; ($D1 + $D2 + $D3 + $D4 + $D5) / 5" | bc)
+
+    # Tier classification
+    # Special rule: D1 < 3 → force bronze
+    if [[ "$D1" -lt 3 ]]; then
+        TIER="bronze"
+        TIER_REASON="D1 Autenticidade < 3 (forced bronze)"
+    elif [[ "$D1" -ge 4 && "$D2" -ge 4 && "$D3" -ge 4 && "$D4" -ge 4 && "$D5" -ge 4 ]]; then
+        TIER="crown_jewel"
+        TIER_REASON="All dimensions >= 4"
+        CROWN_COUNT=$((CROWN_COUNT + 1))
+        OURO_COUNT=$((OURO_COUNT + 1))
+    elif (( $(echo "$AVG_SCORE >= 4.0" | bc -l) )); then
+        TIER="ouro"
+        TIER_REASON="Average >= 4.0"
+        OURO_COUNT=$((OURO_COUNT + 1))
+    elif (( $(echo "$AVG_SCORE >= 3.0" | bc -l) )); then
+        TIER="mixed"
+        TIER_REASON="Average 3.0-3.9"
+        MIXED_COUNT=$((MIXED_COUNT + 1))
+    else
+        TIER="bronze"
+        TIER_REASON="Average < 3.0"
+        BRONZE_COUNT=$((BRONZE_COUNT + 1))
+    fi
+
+    ALL_SCORES+=("$AVG_SCORE")
+    ALL_TIERS+=("$TIER")
+    ALL_NAMES+=("$FNAME")
+
+    # Accumulate per-source report
+    if [[ "$OUTPUT_FORMAT" == "yaml" ]]; then
+        SOURCE_REPORTS+="    - file: \"$FNAME\""$'\n'
+        SOURCE_REPORTS+="      lines: $LINE_COUNT"$'\n'
+        SOURCE_REPORTS+="      tier: \"$TIER\""$'\n'
+        SOURCE_REPORTS+="      tier_reason: \"$TIER_REASON\""$'\n'
+        SOURCE_REPORTS+="      average_score: $AVG_SCORE"$'\n'
+        SOURCE_REPORTS+="      dimensions:"$'\n'
+        SOURCE_REPORTS+="        autenticidade: { score: $D1, checks: [$D1_1, $D1_2, $D1_3, $D1_4, $D1_5] }"$'\n'
+        SOURCE_REPORTS+="        profundidade: { score: $D2, checks: [$D2_1, $D2_2, $D2_3, $D2_4, $D2_5] }"$'\n'
+        SOURCE_REPORTS+="        atualidade: { score: $D3, checks: [$D3_1, $D3_2, $D3_3, $D3_4, $D3_5] }"$'\n'
+        SOURCE_REPORTS+="        unicidade: { score: $D4, checks: [$D4_1, $D4_2, $D4_3, $D4_4, $D4_5] }"$'\n'
+        SOURCE_REPORTS+="        completude: { score: $D5, checks: [$D5_1, $D5_2, $D5_3, $D5_4, $D5_5] }"$'\n'
+    fi
+
+    if [[ "$VERBOSE" == true ]]; then
+        echo "  [VERBOSE] $FNAME: D1=$D1 D2=$D2 D3=$D3 D4=$D4 D5=$D5 avg=$AVG_SCORE tier=$TIER" >&2
+    fi
+done
+
+# ============================================================================
+# AGGREGATION
+# ============================================================================
+if [[ $TOTAL_FILES -gt 0 ]]; then
+    OURO_PCT=$((OURO_COUNT * 100 / TOTAL_FILES))
+else
+    OURO_PCT=0
+fi
+
+# Overall quality
+if [[ "$OURO_PCT" -ge 80 ]]; then
+    OVERALL="excelente"
+elif [[ "$OURO_PCT" -ge 60 ]]; then
+    OVERALL="aceitavel"
+elif [[ "$OURO_PCT" -ge 40 ]]; then
+    OVERALL="parcial"
+else
+    OVERALL="critico"
+fi
+
+# ============================================================================
+# OUTPUT
+# ============================================================================
+if [[ "$OUTPUT_FORMAT" == "json" ]]; then
+    echo "{"
+    echo "  \"source_assessment\": {"
+    echo "    \"directory\": \"$SOURCES_DIR\","
+    echo "    \"date\": \"$DATE\","
+    echo "    \"summary\": {"
+    echo "      \"total_sources\": $TOTAL_FILES,"
+    echo "      \"ouro_count\": $OURO_COUNT,"
+    echo "      \"mixed_count\": $MIXED_COUNT,"
+    echo "      \"bronze_count\": $BRONZE_COUNT,"
+    echo "      \"crown_jewels\": $CROWN_COUNT,"
+    echo "      \"ouro_percentage\": \"${OURO_PCT}%\","
+    echo "      \"overall_quality\": \"$OVERALL\""
+    echo "    },"
+    echo "    \"sources\": ["
+
+    for i in "${!ALL_NAMES[@]}"; do
+        COMMA=""
+        if [[ $i -lt $((${#ALL_NAMES[@]} - 1)) ]]; then
+            COMMA=","
+        fi
+        echo "      {\"file\": \"${ALL_NAMES[$i]}\", \"tier\": \"${ALL_TIERS[$i]}\", \"score\": ${ALL_SCORES[$i]}}$COMMA"
+    done
+
+    echo "    ]"
+    echo "  }"
+    echo "}"
+else
+    # YAML output
+    echo "# Source Assessment Report"
+    echo "# Generated: $DATE"
+    echo "# Directory: $SOURCES_DIR"
+    echo ""
+    echo "source_assessment:"
+    echo "  directory: \"$SOURCES_DIR\""
+    echo "  date: \"$DATE\""
+    echo ""
+    echo "  summary:"
+    echo "    total_sources: $TOTAL_FILES"
+    echo "    ouro_count: $OURO_COUNT"
+    echo "    mixed_count: $MIXED_COUNT"
+    echo "    bronze_count: $BRONZE_COUNT"
+    echo "    crown_jewels: $CROWN_COUNT"
+    echo "    ouro_percentage: \"${OURO_PCT}%\""
+    echo "    overall_quality: \"$OVERALL\""
+    echo ""
+    echo "  sources:"
+    echo -n "$SOURCE_REPORTS"
+    echo ""
+    echo "# Cost: \$0 (zero tokens)"
+    echo "# Time: <${TOTAL_FILES} seconds"
+fi