Education OS Add-on — False Competence Detection & Repair Routing (Almost-Code Canonical)

ENTITY: False Competence
TYPE: Signal–competence divergence
ROLE: A primary hidden fragility mechanism in Education OS

DEFINITION:
False competence exists when a learner can produce correct answers
in familiar patterns (signals) but fails under variation, time pressure,
or transfer—indicating unstable foundations.

CIVOS MAPPING:
Education OS is the time-axis regeneration pipeline.
False competence creates “pipeline illusions”:
- credentials without robustness
- brittle performance at transitions
- increased replacement lag (Φ_in lower quality)
- higher long-run Civλ (capability decay via weak regeneration)

SIGNATURES:
FC.1 high homework completion but test collapse
FC.2 good performance on repeated templates but failure on novel variants
FC.3 careless errors cluster under time pressure (not random noise)
FC.4 performance swings (high variance) with similar topics
FC.5 inability to explain reasoning (bind failure: concept interface weak)
FC.6 rapid forgetting after short intervals (retention instability)

LEARNING PHASE STATES:
P0: cannot execute even basic tasks; breakdown
P1: fragile; works with help or familiar templates; fails under variation
P2: functional; stable under normal practice
P3: robust; stable under variation + time pressure; transfers to new problems

RULE:
False competence is typically:
P2 appearance, P1 reality.

S.EDU.V1  novel-variant success rate (unseen question types)
S.EDU.V2  transfer success rate (same concept, new context)
S.EDU.V3  explanation integrity (can they produce reasoning chain?)

S.EDU.T1  accuracy under timed conditions
S.EDU.T2  error type clustering under time (systematic vs random)
S.EDU.T3  speed–accuracy stability (does speed destroy correctness?)

S.EDU.R1  delayed recall (1 day / 1 week)
S.EDU.R2  forgetting slope (AHD proxy)
S.EDU.R3  rebuild time after break (how fast P2 returns)

S.EDU.F1  primitive mastery checks (core skill atoms)
S.EDU.F2  prerequisite dependency failures (which prerequisite breaks)
S.EDU.F3  bind integrity of definitions (terms used correctly?)

T0: early warning (variance rising)
T1: intervention required (P1 signature appearing)
T2: stop-loss (collapse risk; truncate load immediately)
T3: extinction threshold (foundation pipeline failing; repeated P0 events)

Stop-loss in learning:
If repeated timed failures and retention collapses occur,
reduce scope and rebuild foundations before more exposure.

If failures are:
- execution errors under time → Layer 1/2 (routines + time mechanics)
- consistent concept confusion → Layer 3 (foundations)
- inability to explain → Layer 4 (binds: concept interface)
- frequent destabilisation due to life volatility → Layer 6 (buffers)

Z0: learner execution (skills, habits)
Z1: family support loops (home routines)
Z2: school/tuition system (instruction binds, practice routing)

RULE:
Most learning repair is Z0 + Z2.

TRUNCATE:
- reduce topic spread (narrow scope)
- pause advanced variants temporarily
- reduce workload volatility
- protect sleep and practice time
Goal: create stitching window.

STITCH:
- rebuild primitives (F1)
- repair definition binds (correct meaning and use)
- practice under controlled variation
- reintroduce time pressure gradually
- verify retention over time

ANTI-ILLUSION:
- stop grading-by-volume (homework completion ≠ competence)
- stop “pattern memorisation” as the success criterion
- require explanation binds (reasoning chain)
- require transfer tests before declaring mastery

P3 TESTS:
- stable accuracy on unseen variants
- stable under time pressure
- stable after delay (retention)
- can explain reasoning (bind integrity)
- can transfer concept to new contexts

MALADAPTIVE REPAIRS (δAd⁻):
- teaching shortcuts that improve answers but not foundations
- over-scaffolding that prevents independent execution
- chasing “model solutions” without primitives
- inflating confidence via easy sets; avoiding variation
- increasing workload without buffer repair (burnout → P0)

FAILURE MODE TRACE (schematic):
student learns templates (signal success)
→ novelty or time pressure introduced
→ systematic errors appear (P1 reality)
→ workload continues; no truncation
→ recovery window disappears (stress interval < repair time)
→ retention collapses (AHD rises)
→ repeated P0 events in tests
→ lane access narrows (confidence + options collapse)
→ long-run replacement lag increases (pipeline loss)

Correct repair trace:
stop-loss truncation
→ rebuild primitives + definition binds
→ controlled variation + timed re-entry
→ retention verification
→ P3 robustness restored

CLAIMS (frozen):
- False competence is signal–competence divergence: P2 appearance, P1 reality.
- It is detected by variation, time-pressure, retention, and foundation sensors.
- Repair requires truncation (create window) + stitching (rebuild primitives and binds).
- P3 confirmation requires transfer + time + retention + explanation integrity.
- Signal-only optimisation is maladaptive and increases collapse risk at transitions.