This is the operator hub for RL: what exists, how pieces fit together, and

which docs/commands to use first.

Related maps:

Doc	Focus
AugmentedIntelligenceSystems.md (AugmentedIntelligenceSystems.md)	Whole-app system map
RLCurriculum.md (RLCurriculum.md)	Domain catalog, smoke tests, deployment ladder
RLDomains.md (RLDomains.md)	Expanded domain catalog detail
MathNN.md (MathNN.md)	In-process neural nets + rl net / learners
LifeRL.md (LifeRL.md)	Advisory multi-domain “life” RL
GamingRL.md (GamingRL.md) / GamingRLFPS.md (GamingRLFPS.md)	FPS gaming RL
DrivingRL.md (DrivingRL.md)	Driving RL
KnowledgeRL.md (KnowledgeRL.md)	Knowledge / claim labeling RL
DesktopIconVision.md (DesktopIconVision.md)	Desktop icon classifier (TF SavedModel + LiteRT) for Computer RL
MinecraftPluginTrainingGuide.md (MinecraftPluginTrainingGuide.md)	Minecraft DQN + MathNN + executive plans
CHILD_PSYCHOLOGY_PARENTING_RL.md (CHILD_PSYCHOLOGY_PARENTING_RL.md)	Safety-bounded parenting coaching

Architecture (three layers)

[text]
┌─────────────────────────────────────────────────────────────┐
│  Typed commands / Lua / bridges (operator surface)          │
│  rl … · life rl … · minecraft dqn … · gaming rl … · nn …    │
└───────────────────────────┬─────────────────────────────────┘
                            │
┌───────────────────────────▼─────────────────────────────────┐
│  Domain agents (bespoke loops)                              │
│  MinecraftDQN · GamingRL · DrivingRL · Knowledge · LifeRL   │
│  + Executive Functions plans (high-level steps)             │
└───────────────────────────┬─────────────────────────────────┘
                            │
        ┌───────────────────┼───────────────────┐
        ▼                   ▼                   ▼
┌───────────────┐   ┌───────────────┐   ┌──────────────────┐
│ RLCore /      │   │ RLAgentNet    │   │ Remote / deep    │
│ Algorithms /  │   │ MathNN        │   │ trainer (optional)│
│ Curriculum    │   │ RLNeural      │   │ RLDistributed    │
│ smoke envs    │   │ local DQN…    │   │                  │
└───────────────┘   └───────────────┘   └──────────────────┘

1. Curriculum / smoke layer — safe synthetic RLEnvs; no live actuation.
2. Bespoke domain agents — real observations (Spigot, client, game mod, HUD).
3. MathNN / RLAgentNet — opt-in in-process neural policies (DQN, dueling,

   actor-critic, bandit, hierarchical, multi-head, world-model, reward-model).

Command families (quick map)

Prefix	Role
rl curriculum … / rl sim …	Domain catalog + synthetic smoke
rl real …	Deployment stages (sim → offline → shadow → gated → live)
rl net …	Per-agent MathNN policy registry
rl cartpole / rl dqn	CartPole ODE + MathNN smoke (optional shield/tripwire)
rl life … / life rl …	Advisory Life RL
nn …	MathNN demos, math tutoring RL, architectures
minecraft dqn …	Minecraft agent (observe/train, net, plan, goals)
gaming rl …	FPS gaming RL
driving rl …	Driving RL
knowledge …	Knowledge acquisition RL

Full routing: rl dispatches through _RLFeatures (curriculum, real, net,

cartpole, life). Prefer dry-run / smoke before live control.

In-process neural RL (MathNN)

Enable a neural policy on an agent:

[text]
nn demo
rl net modes
rl net enable minecraft
rl net mode minecraft dueling-dqn
rl net wm minecraft on
rl net curiosity minecraft 0.05
rl net reward minecraft good
rl net q minecraft
rl net loss minecraft
rl net save minecraft checkpoints/minecraft_mathnn.txt

Modes: dqn, dueling-dqn, actor-critic, bandit, hierarchical,

multihead-dueling, worldmodel (dueling + curiosity), reward-model.

Auxiliary (combinable):

[text]
rl net rm <agent> on|off
rl net blend <agent> <alpha>      # α on env reward; (1−α) on reward model
rl net wm <agent> on|off
rl net curiosity <agent> <scale>

CartPole (ODE dynamics + MathNN):

[text]
rl cartpole 50
rl cartpole 50 dueling curiosity shield tripwire

Details: MathNN.md (MathNN.md).

Domain agents

Minecraft DQN

• Observations: Spigot bridge (preferred) or RCON fallback; optional client keys.
• Neural: minecraft dqn net on / rl net enable minecraft.
• Goals: reach / mine / gather / earn.
• Executive Functions: high-level plan steps (wall-guard, net on, status)

  advance each train/eval step when enabled.

[text]
minecraft bridge start
minecraft dqn plan on
minecraft dqn goal mine iron_ore
minecraft dqn plan detail
minecraft dqn net on
minecraft dqn dry-run
minecraft dqn train 1000
minecraft dqn plan status

See MinecraftPluginTrainingGuide.md (MinecraftPluginTrainingGuide.md).

Gaming FPS / Strategy

[text]
gaming rl init
gaming rl dry-run on
gaming rl algorithm dqn
rl net enable gaming_fps
rl net mode gaming_fps actor-critic

See GamingRL.md (GamingRL.md), GameModBridges.md (GameModBridges.md).

Driving

[text]
driving rl status
rl net enable driving
rl net mode driving dueling-dqn

See DrivingRL.md (DrivingRL.md).

Life RL (advisory)

[text]
life rl domains
life rl smoke all 8
life rl net on
life rl suggest focus_blocks
life rl reward good

See LifeRL.md (LifeRL.md). No external side effects from the policy.

Knowledge / parenting / political

Specialized modules with stronger safety text. Prefer synthetic smoke and human

review before anything that could affect real users.

Executive Functions + RL

Executive planning is not a second motor policy. It runs high-level

setup/coaching steps (status, safety toggles, memory/awareness) while domain RL

handles frame-to-frame actions.

Surface	Commands
Generic	plan create …, plan next, plan detail
RSI	rsi agent plan … (sync/advance with RSI goals)
Minecraft	`minecraft dqn plan on\	off\	sync\	next\	status`

Nested minecraft dqn train/step/observe from executive steps is blocked

to avoid re-entrancy during auto-train.

Safe training ladder

• Math smoke: nn demo, rl cartpole 50
• Curriculum smoke: rl curriculum smoke all 8 96
• Domain dry-run: observe / dry-run only (no live keys or RCON write)
• Short train: small episode count; watch status/loss
• Checkpoint: rl net save / domain save
• Eval gate: promote only if metrics improve
• Live: deployment stages via rl real stage …

Never jump from cold start to live desktop, medical, legal, finance, or public

multiplayer bots.

RSI and Lua

RSI scripts get MathNN + executive Lua globals, for example:

[lua]
ai_nn_modes()
ai_nn_policy_enable(“life”, “dueling-dqn”, 16, 8)
rsi_sync_executive_plan(false)
rsi_advance_executive_plan(1)
executive_status()
ai_rl_status()

RSI cycles can seed and advance executive plans while the CommandRouter runs

perception/reasoning commands. See RSI scripts under lua_scripts/RSI-*.lua.

File layout (code)

Area	Files
Substrate	RLCore, RLAlgorithms, RLExplore, RLFeatures, RLCurriculumOrchestrator, RLDeployment, RLReal, RLRealEnvs
Neural	MathNN*, RLNeural, RLAgentNet, MathNNAgentBridge, MathNNIntegration
Domains	MinecraftDQN, GamingRL, DrivingRL, KnowledgeAcquisitionRL, LifeRL, ChildPsychologyParentingRL, PoliticalRL
Planning	Executive Functions

Troubleshooting

Symptom	Check
rl net not recognized	Ensure rl routes via _RLFeatures (not curriculum-only).
Net resets every tick	Agent must pass a stable rl_mode to EnsureConfigured (Minecraft uses dueling-dqn).
No learning	Bad state/reward/terminal wiring first; then LR/epsilon.
Minecraft walks into walls	Wall-guard, bridge collision features, dry-run; not only network size.
Executive flips train off	Mode-changing steps are blocked; use plan detail to inspect steps.
Life RL “does nothing real”	By design — advisory only.

Version note

Documentation updated for the MathNN multi-learner registry, Life RL, Minecraft

executive plan hooks, RSI executive integration, and unified rl command

routing (2026).

---

Filed under: Uncategorized - @ July 18, 2026 6:50 am