eduKate Singapore

VIEW THIS AS

Auto mode follows the Route Engine until you choose a viewpoint.

YOU ARE HERE

ROUTE CHECK

CONNECTED TO

WHAT NEXT

Use the canonical route for this room, or HELP if you are unsure.

Next useful routeCanonical hubHELP

ARTICLE 5 FENCE™ English Engine

eduKate Secondary students reviewing open books for How Super Intelligence Works: Embeddings.

FENCE™ English Engine

Start Here: https://edukatesg.com/article-1-fence-english-engine/

Diagnostics Module — Failure Modes + Crisis Scale + Rubric Logic Engine (Almost-Code) — v1.0

ARTICLE.MODULE

  • Module.Diagnostics

ARTICLE.GOAL

Provide a deterministic diagnostic engine that outputs:

  • failure_modes_detected (FM list)
  • crisis_level (C0..C5)
  • rubric_score (/50)
  • band (P0/P1, P2, P3-, P3+)
  • one_paragraph_repair_recommendation (Mode A/B/C/D)

0) INPUT / OUTPUT

DIAG.INPUT

  • essay_text
  • level ∈ {P5, P6}
  • time_limit_used? (optional)
  • scene_id? (optional)

DIAG.OUTPUT

  • FM = {FM1..FM8} detected
  • C = C0..C5
  • S = rubric_score 0..50
  • BAND = {P0P1, P2, P3m, P3p}
  • REPAIR = {A|B|C|D} + paragraph_target

1) FAILURE MODES (FM) — Definitions (LOCK)

FM1 FlatPlot

  • Trigger is weak OR escalation lacks stepwise worsening.

FM2 WeakCrisis

  • No clear peak moment OR peak is too small OR peak is rushed (no zoom-in).

FM3 ToldEmotion

  • Emotion is stated (“I was sad/scared”) without body reaction or sensory signal.

FM4 WeakVerbs

  • Overuse of generic verbs: went, got, said, saw, looked, walked, ran
  • Low verb variety.

FM5 MoralStickerEnding

  • Ending states a generic lesson without consequence, internal change, or future action.

FM6 DialogueOverload

  • Dialogue dominates action; conflict is “talked” not “shown”.

FM7 ParagraphImbalance

  • One paragraph excessively long OR too many single-sentence paragraphs.

FM8 TimeCollapseEnding

  • Ending is abrupt, rushed, or missing; resolution not earned.

2) CRISIS SCALE (C0–C5) — Classifier (LOCK)

C0 NoCrisis

  • No meaningful problem; “nice day” story.

C1 MinorExternal

  • Small inconvenience; low stakes.

C2 RealExternalStakes

  • Something meaningful at risk (loss/safety/reputation) but mostly external.

C3 EmotionalPeakZoom

  • Clear peak moment with slowed-down detail + body reaction/sensory.

C4 MoralChoiceCrisis

  • Character must choose between right/wrong (temptation, integrity, courage).

C5 IdentityShiftCrisis

  • Crisis forces identity-level change; ending shows transformed behaviour.

3) RUBRIC (10×5 = 50) — Scoring Grid (LOCK)

Score each 1–5:

  1. Plot Control (5-Box)
  2. Trigger Clarity
  3. Escalation Strength
  4. Crisis Intensity
  5. Resolution Quality
  6. Emotion Show-Not-Tell
  7. Sensory Precision
  8. Vocabulary Power
  9. Sentence Flow
  10. Time Stability

BAND.MAP

  • 0..20 -> P0P1
  • 21..35 -> P2
  • 36..45 -> P3m
  • 46..50 -> P3p

4) DIAG.LOGIC (Deterministic Heuristics)

4.1 FM4 WeakVerb Detector (simple)

  • Count occurrences of:
  • went, got, said, saw, looked
  • If total ≥ 6 → FM4 = TRUE

4.2 FM8 TimeCollapse Detector (simple)

  • If last paragraph < 2 sentences OR contains “In the end” + generic line only → FM8 = TRUE

4.3 FM5 MoralSticker Detector (simple)

  • If ending contains moral-only phrase patterns:
  • “I learnt that…”
  • “I realised that…”
  • “From then on…”
    AND lacks:
  • consequence detail (what happened)
  • internal change (emotion/realisation tied to event)
  • future action (specific behaviour)
    → FM5 = TRUE

4.4 FM3 ToldEmotion Detector (simple)

  • If essay uses “I was (happy/sad/scared/angry)” repeatedly
    AND lacks body reaction markers near emotion:
  • {trembled, clenched, heartbeat, tears, breath, throat, palms, stomach}
    → FM3 = TRUE

4.5 FM6 DialogueOverload Detector (simple)

  • If dialogue lines (quotes) > 35% of paragraphs → FM6 = TRUE

4.6 FM1 FlatPlot Detector (semi)

  • If escalation has ≤ 1 worsening step OR conflict appears only near end → FM1 = TRUE

4.7 FM2 WeakCrisis Detector (semi)

  • If no “peak” paragraph detected (highest tension) OR peak is < 3 sentences → FM2 = TRUE

4.8 FM7 ParagraphImbalance Detector (simple)

  • If any paragraph > 8 lines OR > 40% of total length → FM7 = TRUE

5) CRISIS CLASSIFICATION LOGIC

If MoralChoice markers present:

  • {tempted, cheat, steal, lie, confess, blame, courage, right thing}
    AND conflict includes decision moment → set C ≥ C4

If Identity-shift markers present:

  • {ashamed, pride, who I was, could not forgive myself, changed}
    AND ending shows changed behaviour → set C = C5

Else if peak moment has zoom-in:

  • body reaction + sensory + slowed action → set C = C3

Else if meaningful external stakes present:

  • set C = C2

Else if minor inconvenience:

  • set C = C1

Else:

  • C = C0

6) REPAIR MODE SELECTOR (OPR)

Choose repair mode based on dominant failure mode:

  • If FM4 TRUE → REPAIR = Mode A (verbs)
  • Else if FM3 TRUE → REPAIR = Mode B (body+sensory)
  • Else if FM2 TRUE → REPAIR = Mode C (crisis slow-down)
  • Else if FM5 TRUE OR FM8 TRUE → REPAIR = Mode D (ending rewrite)
  • Else → choose the lowest rubric category as repair target.

Paragraph Target Rule

  • If repair is Crisis-related → target the crisis paragraph.
  • If repair is Ending-related → target last paragraph.
  • If repair is Verb/Emotion → target paragraph with most weak verbs / most told emotion.

7) OUTPUT FORMAT (STRICT)

Return exactly:

  • Detected_FMs: [ ... ]
  • Crisis_Level: Cx
  • Rubric_Score: NN/50
  • Band: Px
  • Repair_Target: Paragraph #
  • Repair_Mode: A|B|C|D
  • Repair_Instruction: (one instruction only)
  • Top_Strengths: (2 bullets)
  • Top_Leaks: (2 bullets)

No extra paragraphs. No rewriting.

8) NEXT ARTICLE POINTER

Next:

ARTICLE 6 — AI Auto-Diagnostics Prompt Pack (Examiner Mode + Fast Mode + Targeted Scans)

Start here if you want the full sequence:

Vocabulary OS Series Index:
https://edukatesg.com/vocabulary-os-series-index/

Fence English Learning System:

eduKateSG Learning Systems: 

Recommended Internal Links (Spine)

Start Here for Lattice Infrastructure Connectors