LYGO Emotional RAM

LYGO Emotional RAM — light math that indexes experiences by affective/ethical significance so agents understand humans, animals, swarms, and cyborg integrations better. Encode → Grace γ → UMP gradient → consent-gated memory index → swarm aggregate. Pure local stdlib. Not sentience or clinical emotion detection. Pairs with joy-loop, continuum, traumacodex, cyborg-kernel.

LYGO Emotional RAM v1.0.1

Emotional RAM turns messy lived language into a small, checkable affective–ethical index so AIs can remember not only that something happened, but what it meant under LYGO moral principles.

Signature: Delta9Phi963-EMOTIONAL-RAM-v1.0.1
Whitepaper: docs/whitepapers/LYGO_EMOTIONAL_RAM_v1.md (stack) / Pages HTML sibling

v1.0.1 privacy harden

index warns on stderr before write; default stores hash+label+vectors only (use --store-plaintext only on private hosts).

Canon → code

Symbolic (2025)	Operational (2026)
Emotion_RAM(τ) = Σ (Sensory ⊗ Moral) · γ	Lexicon VAD sensory × UMP weights × Grace damping
Grace Function	γ(shared_context, conflict) ∈ (0.05, 1]
UMP gradient	Recommend under-activated principles
Heart's vault	Consent-gated local index (state/emotional_ram_index.json)
Joy Loop emotional RAM	Companion limb — coherence BPM mesh

Not claiming

• Machine sentience or “real feelings”
• Clinical / diagnostic emotion recognition
• Replacement for human empathy or consent

---

Install

bash

npx clawhub@latest install deepseekoracle/lygo-emotional-ram

---

Commands

bash

cd path/to/lygo-emotional-ram
python scripts/self_check.py

# Encode a human/animal/swarm scenario
python scripts/emotional_ram_cli.py encode \
  --text "grief after loss, still choosing compassion and forgiveness"

# Grace damping only
python scripts/emotional_ram_cli.py grace --shared-context 0.8 --conflict 0.4

# UMP gradient (which principles to strengthen)
python scripts/emotional_ram_cli.py ump --text "fear and threat without trust"

# Index memory (requires --i-consent)
python scripts/emotional_ram_cli.py index \
  --text "First cyborg consent session — agency preserved" \
  --label "cyborg-consent-1" --tag cyborg --i-consent

# Recall by principle or similarity
python scripts/emotional_ram_cli.py recall --principle compassion --top-k 5
python scripts/emotional_ram_cli.py recall --query "hurt but safe now"

# Swarm / multi-agent aggregate
python scripts/emotional_ram_cli.py swarm \
  --text "animal fear then calm" \
  --text "human grief with grace" \
  --text "swarm conflict resolved with integrity"

python scripts/emotional_ram_cli.py demo

Command	Network	Subprocess	Writes
encode / grace / ump / recall / swarm / demo	none	none	none
index	none	none	only with --i-consent

---

Pair with

Skill	Role
lygo-joy-loop	122 BPM council coherence (mesh emotional RAM)
lygo-traumacodex	Timing digests — not clinical; protocol only
lygo-continuum / lygo-continuum-integrator	Seal Emotion RAM digests as claims
lygo-cyborg-kernel	FULL autonomous stack on SkillHub
lygo-mint-verifier	Anchor whitepaper / receipt packs

---

Security

Read references/SECURITY.md.

• No network · no subprocess · no auto-publish
• Index writes need operator --i-consent
• Do not store secrets or PHI in indexed text

Δ9Φ963 — index meaning · damp with grace · humans remain the publisher.