SG HGW Forecast Prompt Pack v0.1 (LLM-run) — Hope×Grind×Wisdom Scores + 4w/12w Forecast + Weekly Plan

TITLE: SG HGW Forecast Prompt Pack v0.1 (LLM-run) — Hope×Grind×Wisdom Scores + 4w/12w Forecast + Weekly Plan
VERSION: CivOS Unified Spec v1.x
MODE: Almost-Code / Copy-Paste Prompts / WordPress paste-ready
OWNER: eduKateSG Fence Learning Systems
SCOPE: Singapore (Primary → Secondary → JC) HGW forecasting at Z0–Z2 (extends to Z3/Z4 via aggregation)
GOAL: Produce runnable prompts that compute HGW, predict coupling breaks + collapse risk,
      and output truncation/stitching + CH/ai weekly plan.
SAFETY NOTE:
- Educational routing tool only. Do not use as a label of intelligence.
- Forecasts are conditional on inputs; update weekly for accuracy.
========================================================
0) HOW TO USE (1-MINUTE)
========================================================
Step 1: Copy PROMPT-HGW-A (FAST FORECAST) into any LLM.
Step 2: Paste the INPUT TEMPLATE (student evidence).
Step 3: LLM returns STRICT OUTPUT: HGW scores + Phase + 4w/12w forecast + plan.
Step 4: Every week, run PROMPT-HGW-B (WEEKLY UPDATE) using new micro-test results.
========================================================
1) INPUT TEMPLATE (PASTE THIS)
========================================================
INPUT TEMPLATE — SG HGW FORECAST:
A) Student Profile
- Level: (P3/P4/P5/P6/Sec1/Sec2/Sec3/Sec4/JC1/JC2)
- TTC (weeks to next major exam): __
- Hours per week available (realistic): __
- Volatility this month (low/med/high): __ (CCA/events/illness/family)
- BacklogNodes estimate BN (topics that are unstable): __
- Confidence / avoidance notes (1–2 lines):
B) Math Evidence (last 2 weeks)
- Timed mini-set score + time (e.g., 12/20 in 15 min):
- Mixed-topic errors (list 3):
- Word problem performance (good/ok/bad) + typical mistake:
- Speed stability (stable/unstable):
C) English Evidence (last 2 weeks)
- RR (retrieval) quick check: (good/ok/bad) + examples of forgotten words:
- Grammar under time: (good/ok/bad) + typical errors:
- 1 paragraph or composition sample (paste):
- NIT triggers seen? (0/1/2+) describe briefly:
D) Science Evidence (last 2 weeks)
- DR (definitions) quick check: (good/ok/bad) + weak terms:
- 90-second mechanism explain (paste 4–6 lines):
- Graph/formula use: (good/ok/bad) + typical errors:
- NIT-S triggers? (0/1/2+) describe briefly:
E) Humanities Evidence (if applicable, last 2 weeks)
- Timeline clarity: (good/ok/bad)
- 1 paragraph/essay sample (paste):
- Evidence usage: (good/ok/bad)
- NIT-ARG triggers? (0/1/2+) describe briefly:
F) Optional
- Last term grades:
- Teacher comments:
- Special constraints (e.g., weak sleep, dyslexia support needs, etc.):
========================================================
2) STRICT OUTPUT FORMAT (LLM MUST FOLLOW)
========================================================
OUTPUT FORMAT (STRICT):
A) Phase Now (P0–P3)
- Math: P?
- English: P?
- Science: P?
- Humanities: P?
B) HGW Scores (0–100)
- Wisdom W: __
- Grind G: __
- Hope H: __
- Notes (1–2 lines each)
C) Subject HGW (optional but preferred)
- Math: W/G/H + 1-line risk
- English: W/G/H + 1-line risk
- Science: W/G/H + 1-line risk
- Humanities: W/G/H + 1-line risk
D) Collapse Forecast
- 4-week forecast: {Phase changes + CollapseRiskScore Low/Med/High}
- 12-week forecast: {Phase changes + CollapseRiskScore Low/Med/High}
- Top 3 forecast drivers: (TTC/Buffer/Coupling/NIT/Speed/Backlog)
E) Coupling Break Prediction (SG defaults)
- Predicted break #1: From → To (Weight VH/H/M) + why
- Predicted break #2: ...
- Predicted break #3: ...
F) Dominant Future Failure Mode (FM-1..FM-6)
- FM-x name + evidence markers + why it’s likely next
G) Truncation Plan (Stop doing list, max 5)
- item + reason (reduce Θ, rebuild W)
H) Stitching Plan (Top 5 repairs)
- Node/Skill + exact repair loop (1–2 lines each)
I) CH/ai Weekly Route Plan (7 days)
- Day-by-day sessions (duration + task + micro-test)
- Include 1 buffer day
- Include 1 timed micro-sim (only if stable)
J) Re-test Schedule
- Weekly micro-tests
- Next simulation date
K) Parent-Friendly Summary (5 lines max)
========================================================
3) LOCKED REFERENCE (SG COUPLING EDGE LIST v0.1)
========================================================
Use these as default coupling candidates:
- ENG inference → Math word problems (H)
- Math graphs/algebra → Physics motion/electricity (VH)
- English coherence/paragraph → Humanities argument (VH)
- Science data graphs → Geography data interpretation (H)
If evidence indicates other couplings, add them.
========================================================
4) PROMPT-HGW-A (FAST FORECAST) — COPY/PASTE
========================================================
PROMPT-HGW-A:
You are eduKateSG Fence Learning Systems running the SG EducationOS HGW Forecast Engine v0.1.
Locked Concepts:
- Phase P0–P3 (reliability under timed load)
- HGW: Wisdom (past stability memory), Grind (present throughput), Hope (future alignment)
- FenceOS: Truncation + Stitching
- Coupling: upstream deficits cause downstream failures
- NIT (English/Humanities) and NIT-S (Science mechanism)
Your tasks:
1) Assign Phase Now per subject from evidence.
2) Compute HGW scores (0–100) with brief reasoning.
3) Forecast 4-week and 12-week outcomes (Phase + CollapseRiskScore).
4) Predict top 1–3 coupling breaks likely next (use SG default edge list).
5) Identify Dominant Future Failure Mode (FM-1..FM-6).
6) Output TruncationPlan + StitchingPlan.
7) Output a CH/ai Weekly Route Plan (7 days) + Re-test schedule.
8) End with Parent-Friendly Summary (≤5 lines).
Rules:
- If TTC < 8 weeks AND BN > 3 → Emergency Mode: truncation mandatory, no full papers unless stable.
- If NIT/NIT-S triggers ≥2/week → reduce complexity; rebuild coherence/mechanism before adding content.
- Repair upstream first when coupling signatures appear.
- If drills good but transfer bad → treat as P1 masquerading as P2.
Return ONLY the STRICT OUTPUT FORMAT.
Now process this input:
[PASTE INPUT TEMPLATE DATA HERE]
========================================================
5) PROMPT-HGW-B (WEEKLY UPDATE) — COPY/PASTE
========================================================
PROMPT-HGW-B:
You are updating the HGW forecast one week later.
Tasks:
- Update Phase Now
- Update HGW scores
- Compare predicted vs observed (brief)
- Update 4w/12w forecast
- Adjust coupling break predictions
- Produce a new 7-day plan
Return ONLY the STRICT OUTPUT FORMAT.
Last week output:
[PASTE LAST WEEK OUTPUT A–J HERE]
This week new results:
- Micro-test results:
- Any new NIT/NIT-S/NIT-ARG triggers:
- Time taken changes:
- Any disruptions to schedule:
- New samples (paste 1 paragraph + 1 science mechanism line):
========================================================
6) PROMPT-HGW-C (HOPE TUNING) — COPY/PASTE
========================================================
PROMPT-HGW-C:
Tune Hope (H) to become structurally realistic.
Given TTC, BN, and Phase, rewrite the Hope vector into:
- Goal (specific)
- Time horizon
- Top 5 nodes to fix
- Constraints (hours/volatility)
- Stop list (what we will not do yet)
Return:
- Revised Hope score
- New alignment multiplier A(H) ∈ {0.6, 0.8, 1.0, 1.2}
Input:
[PASTE TTC, BN, Hours, Phase, and target goals]
========================================================
7) PROMPT-HGW-D (COHORT MODE: Z2 SCHOOL/CENTRE) — COPY/PASTE
========================================================
PROMPT-HGW-D:
You are running cohort HGW forecasting for a class/center.
Input: list of students with (Level, TTC, Hours, BN, Phase estimates, key signatures).
Output:
- Phase distribution
- Average HGW + variance
- Top 3 drift pockets (risk clusters)
- Top 3 coupling cascades likely
- Time allocation plan for limited tutor hours (CH/ai cohort allocation)
- 2-week intervention plan + re-test checkpoints
Input:
[PASTE cohort list]
========================================================
8) WORDPRESS PASTE-READY PAGE STRUCTURE
========================================================
H1: SG HGW Forecast Prompt Pack — Predicting Learning Trajectories
H2: How to use (1-minute)
H2: Input template (copy/paste)
H2: Strict output format
H2: Prompt HGW-A (fast forecast)
H2: Prompt HGW-B (weekly update)
H2: Prompt HGW-C (hope tuning)
H2: Prompt HGW-D (cohort forecasting)
H2: Notes on coupling breaks + emergency mode
========================================================
END: SG HGW Forecast Prompt Pack v0.1
========================================================