katopz/katgpt-rs / .agents/skills/research
research
VerifiedResearch workflow for distilling ML/AI papers into modelless inference primitives, freeze/thaw runtime patterns, latent-space operations, AND model-based training plans across the multi-repo stack. Use when reading arxiv papers, deciding which repo a paper belongs in, creating .research/ notes or .plans/ files, implementing modelless inference primitives, or routing training-vs-inference insights. Enforces the commercial strategy (public engine / private runtime / private chain / private neuron-db / private training / private SDK facade / private product-domain / public inference-substrate riir-infer / public decision-serving riir-reflex), three-track system (modelless inference + self-adaptive runtime + model-based trained weights — all three exist across the stack, not just riir-train), latent-to-latent preference, and freeze/thaw-over-fine-tuning rule. Every task ends with a mandatory Claude verdict ping-pong (§5) before any file is committed.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 80.0 KB |
| latent_raw.md | markdown | 1.6 KB |
| templates.md | markdown | 4.2 KB |
| vocab.md | markdown | 31.9 KB |
latent_raw.md
View on GitHub# Latent vs raw space rules (critical for game AI)
- **Physical** (position, velocity, HP, wallet): MUST be raw exact. Deterministic replay, quorum sync, anti-cheat require bit-identical reconstruction.
- **Semantic** (emotion, mood, curiosity, style, habit): SHOULD be latent via dot-product + sigmoid onto learned direction vectors.
- **Social** (encounters, relationships, factions): SHOULD produce KG triples from latent/embedding proximity, not raw coordinate distance.
**Sync boundary:** if data flows through `SyncBlock → ChainConsensus` quorum → Cold tier, it MUST be raw + deterministic. If consumed locally (emotion projection, shard retrieval, consolidation sleep-cycle), it SHOULD be latent. Bridge functions (raw→latent projection, latent→raw scalar clamp) MUST be zero-alloc, gateable, sync-invariant.
**KG triple emission:** semantic encounters → KG triple from latent similarity. Physical events → TxDelta with raw values, NOT KG triple. Never substitute latent embedding for raw position in anti-cheat.
**Spatial cognition (two-brain model):** info brain = real `MapPos` (synced, ground truth). Think brain = per-NPC `SpatialBelief` (zone KG triple + stale last_known_pos, fog-of-war gated, NOT synced). Bridge is one-way: real position → belief update only when within `visible_radius`. Confidence decay: `sigmoid(-λ * (current_tick - last_observed_tick))`. Two brains MUST exist independently — divergence is emergent, not a bug.
Moved verbatim out of `SKILL.md` on 2026-10-28 to keep that file under the skill-size gate. The global agent rules (`~/.claude/CLAUDE.md` §"Latent vs Raw Space Rules") are the canonical, fuller copy.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |