A Recovery Mode (RM) is a primary repair profile matched to a failure signature.
Every run selects exactly one RM to avoid advice sprawl and to make retesting valid.
Each RM includes:
- When to use it (DLR pattern)
- What to do first
- What to stop
- Minimum probes (retest thresholds)
RM-00 — Stabilise the Loop (Emergency Containment)
Purpose: Stop runaway collapse when Load is high and Repairability is failing.
Use when (DLR):
- L ≥ 4 and R ≤ 2 in any OS
- or any OS shows accelerating decline week-to-week
Do first:
- Reduce load immediately (cut scope, pause non-essentials)
- Protect truth channels (stop data suppression)
- Establish a daily/weekly correction cadence
Stop:
- expansion, new reforms, new projects
- punishment of bad news
- “more output” demands without repair capacity
Probes:
- R increases by +1 within 14–30 days
- error reporting volume rises (truth returning)
- burnout/attrition stabilises (no further rise)
Education OS Recovery Modes (Capability Engine)
RM-01 — Foundation Rebuild Mode
Purpose: Restore Depth when the system is brittle.
Use when (DLR):
- D ≤ 2 with unstable performance
- memorisation without understanding
Do first:
- rebuild prerequisites (smallest units → combine)
- slow down content coverage
- enforce explain-back and concept transfer
Stop:
- volume-based drilling as the primary strategy
- skipping prerequisites
- testing without correction
Probes:
- explain-back accuracy ≥ 80%
- novel problem success improves by +20% in 4–8 weeks
- error type shifts from conceptual to minor execution
RM-02 — Transfer Repair Mode
Purpose: Restore transfer when learning does not convert into real performance.
Use when (DLR):
- T (or functional transfer) ≤ 2 while D may be moderate/high
- exam performance collapses under novelty/stress
Do first:
- varied practice (context switching)
- retrieval under time pressure
- teach “why” and “when” rules apply
Stop:
- same-format practice only
- over-scaffolding that removes thinking
Probes:
- novel-context tasks pass rate rises weekly
- timed application improves without extra hints
- student can self-generate examples
RM-03 — Load Pressure Relief Mode
Purpose: Reduce overload so learning can repair.
Use when (DLR):
- L ≥ 4 and motivation/health degrading
- burnout symptoms, avoidance, panic loops
Do first:
- reduce simultaneous demands (one priority track)
- restore sleep and recovery
- shorten cycles: learn → test → repair
Stop:
- stacking tuition + CCAs + heavy homework without slack
- punishment-based motivation
Probes:
- fatigue markers decrease in 2–3 weeks
- attendance consistency improves
- R rises (self-correction resumes)
RM-04 — Meaning & Identity Re-anchor Mode
Purpose: Restore intrinsic drive and self-worth (energy over time).
Use when (DLR):
- R collapsing due to shame, cynicism, “I’m not smart” identity
- effort stops despite feasible workload
Do first:
- redefine success as trajectory, not status
- build micro-wins and visible progress graphs
- reconnect learning to purpose and belonging
Stop:
- humiliation, comparison framing
- “lazy” labels (they destroy R)
Probes:
- voluntary practice time increases
- self-report hope/agency rises
- error correction happens without prompting
RM-05 — Talent Alignment Mode
Purpose: Reduce waste and restore growth by matching pathway to hardware.
Use when (DLR):
- persistent high load with low progress despite effort
- chronic mismatch (wrong pace, wrong modality, wrong track)
Do first:
- identify strength geometry (verbal/spatial, speed/depth)
- redesign pathway (tools, pacing, modality)
- remove unnecessary friction
Stop:
- forcing uniform methods
- one-size-fits-all “discipline fixes”
Probes:
- same effort produces faster learning
- stress drops while performance rises
- consistency improves
Governance OS Recovery Modes (Steering Engine)
RM-06 — Truth Restoration Mode
Purpose: Restore measurement integrity and reality contact.
Use when (DLR):
- Governance R ≤ 2 with hidden errors, censored data, or fake success
- policy decisions disconnected from ground truth
Do first:
- independent audits
- publish correction metrics
- protect whistleblowing channels
Stop:
- punishment for bad news
- metric-only success narratives
Probes:
- error discovery increases initially (good sign)
- audit variance shrinks over time
- policy changes track measured reality
RM-07 — Incentive Realignment Mode
Purpose: Fix Goodhart loops and gaming.
Use when (DLR):
- outcomes degrade while metrics improve
- reward systems produce perverse behavior
Do first:
- redefine metrics around real outcomes
- tie rewards to transfer and long-horizon results
- add anti-gaming constraints
Stop:
- single-metric targets
- prestige-only reward loops
Probes:
- metric/outcome correlation improves
- gaming behaviors drop
- long-horizon outcomes rise
RM-08 — Legitimacy Rebuild Mode
Purpose: Restore trust and voluntary compliance.
Use when (DLR):
- L high in governance + rising distrust/polarisation
- coercion increasing to maintain order
Do first:
- fairness repair (transparent rules)
- accountability mechanisms
- consistency and visible correction
Stop:
- arbitrary enforcement
- hypocrisy (destroys legitimacy fastest)
Probes:
- trust indicators rise
- compliance improves without force
- conflict incidents decline
RM-09 — Feedback Loop Reopening Mode
Purpose: Restore self-correction capacity.
Use when (DLR):
- repeated known failures with no changes
- “nothing can be done” institutional paralysis
Do first:
- create mandatory review cycles
- implement small experiments with retests
- make correction non-punitive
Stop:
- blame rituals
- endless planning without implementation
Probes:
- correction cadence becomes regular
- repeat failures decrease
- R improves by +1 within 1–2 cycles
Production/Technology OS Recovery Modes (Power Amplifier)
RM-10 — Resilience & Redundancy Mode
Purpose: Reduce fragility and single points of failure.
Use when (DLR):
- high shock sensitivity, supply breaks, cascading outages
Do first:
- add redundancy in critical paths
- diversify suppliers/energy sources
- improve maintenance schedules
Stop:
- cost-cutting that removes buffers
- over-optimization for efficiency only
Probes:
- downtime decreases
- recovery time after shocks improves
- failure cascades reduce
RM-11 — Maintenance & Renewal Mode
Purpose: Reverse entropy and under-maintenance.
Use when (DLR):
- D moderate but output falling due to decay
- infrastructure aging and breakdowns increasing
Do first:
- inventory critical assets
- fund maintenance before expansion
- rebuild institutional maintenance discipline
Stop:
- expansion projects that increase maintenance debt
Probes:
- backlog shrinks steadily
- failure frequency declines
- costs stabilise
RM-12 — Governance–Tech Alignment Mode
Purpose: Prevent tech outrunning governance (safety/ethics/security).
Use when (DLR):
- rapid deployment causing social harm or security risks
- regulatory lag producing chaos
Do first:
- align incentives for safe deployment
- build guardrails and audits
- phase rollouts with retests
Stop:
- “move fast and break society” logic
Probes:
- incident rate declines
- public trust stabilises
- adoption continues without backlash
Constraint OS Recovery Modes (Reality Boundary Layer)
RM-13 — Load Reduction & Demand Shaping Mode
Purpose: Reduce pressure when constraints tighten.
Use when (DLR):
- constraint load rising faster than adaptation capacity
- scarcity/overshoot signals increasing
Do first:
- cut waste
- reduce non-essential consumption
- redesign incentives toward efficiency
Stop:
- denial narratives
- growth-at-any-cost under scarcity
Probes:
- consumption intensity drops
- scarcity pressure stabilises
- regeneration indicators improve
RM-14 — Substitution & Feasible Set Expansion Mode
Purpose: Shift to alternatives when a constraint binds.
Use when (DLR):
- specific bottleneck resource/energy bind
Do first:
- invest in substitutes
- redesign systems to reduce dependency
- accelerate innovation pathways with safe governance
Stop:
- doubling down on depleted inputs
Probes:
- dependency ratio declines
- cost volatility reduces
- stability improves
RM-15 — Regeneration & Restoration Mode
Purpose: Restore the boundary conditions (ecology, resources).
Use when (DLR):
- overshoot beyond regeneration capacity
Do first:
- protect regeneration systems (water, soil, ecosystems)
- redirect production toward restoration
- enforce long-horizon constraints
Stop:
- extraction that exceeds regeneration
Probes:
- regeneration indicators rise
- ecological debt decreases
- long-term sustainability improves
Selection Rule (Mandatory)
Pick exactly one RM based on the lowest Repairability (R) and highest Load (L).
If multiple are failing:
- choose RM-00 Stabilise the Loop first
- then rerun DLR after stabilisation
