diff --git a/docs/architecture/aetheria-cold-wake-training-fixture.md b/docs/architecture/aetheria-cold-wake-training-fixture.md index 3918417..e2430bf 100644 --- a/docs/architecture/aetheria-cold-wake-training-fixture.md +++ b/docs/architecture/aetheria-cold-wake-training-fixture.md @@ -3,15 +3,7 @@ Cold Wake is useful to Ghostlight as a grounded historical pressure cooker, not as the active product target. -The deleted `aetheria-cold-wake-mvp-scenario.md` brief had the right social -machine in it: Ganymede corridor law, thermal stealth panic, Navigator rescue -obligation, insurer category failure, shadow cognition, and institutions forced -to translate the same ambiguous vessel through incompatible moral economies. -The bad part was the label. "MVP scenario" quietly points the repo back toward -building a game slice when the current target is projection, content generation, -and dialogue scaffolding. - -So the revived shape is narrower: +The revived shape is narrow: - use Cold Wake as an authored historical Aetheria flashpoint - distill it into lore grounding digests @@ -43,11 +35,15 @@ The first machine-readable digest is: - `examples/lore-grounding/cold-wake-panic.ganymede-corridor.json` -It is marked `draft` because the next move is not to canonize a whole plotted -incident. The next move is to choose one room inside the crisis and build a -projection fixture from it. +It can ground Cold Wake projection, responder, coordinator, and branch compiler +artifacts, including: -## Candidate First Rooms +- `examples/ink/cold-wake-branch-and-fold.v0.ink` +- `examples/ink/cold-wake-branch-and-fold.v0.training.json` +- `examples/coordinator/cold-wake-branch-and-fold.v0.json` +- `experiments/cold-wake-story-lab/cold-wake-branch-and-fold-clean-run.v0.md` + +## Useful Rooms - PSC risk tribunal freezes a quiet-running vessel's corridor status while a Navigator route steward argues rescue cannot wait for category repair. diff --git a/docs/architecture/aetheria-lore-grounding-architecture.md b/docs/architecture/aetheria-lore-grounding-architecture.md index 2127cc1..9528cd6 100644 --- a/docs/architecture/aetheria-lore-grounding-architecture.md +++ b/docs/architecture/aetheria-lore-grounding-architecture.md @@ -374,18 +374,13 @@ Blocking gaps should be patched in AetheriaLore first. Safe stubs must be tracked in the digest or sidecar and revisited before promotion. Deferred gaps should stay out of prompt text so they do not harden by accident. -## Immediate Plan +## Training Boundary -1. Keep the Cat/Oz fixture as an Elysium procedural mechanics fixture. -2. Add a lore grounding digest format. Done. -3. Choose one authored historical flashpoint from the Aetheria archive. -4. Build the first `historical_grounded` agent-state fixture from that slice. -5. Produce reviewed projection examples from that fixture. -6. Create the first `future_branch` fixture from authored post-Elysium concepts - after the coordinator artifact schema exists. -7. Compare those examples against Cat/Oz to separate grounded historical - pressure from procedural Elysium branch generation. - -Ground the machine in the past. Then generate possible futures under -supervision, with branch labels sharp enough that infinity does not show up +Ground the machine in the past, then generate possible futures under +supervision with branch labels sharp enough that infinity does not show up wearing one nametag. + +Historical fixtures teach single-history social pressure. Future-branch +fixtures teach conditional canon. Cat/Oz-style procedural fixtures teach +runtime mechanics. Those lanes can compare against each other, but they should +not collapse into one unlabeled corpus. diff --git a/docs/architecture/ink-branching-scenes.md b/docs/architecture/ink-branching-scenes.md index d63f717..74990e0 100644 --- a/docs/architecture/ink-branching-scenes.md +++ b/docs/architecture/ink-branching-scenes.md @@ -23,7 +23,7 @@ cultural, memory, relationship, and consequence rationale. ## Runtime Shape -The near-term authoring path is: +The artifact flow is: ```text agent state + scene state + lore grounding diff --git a/docs/architecture/projection-distillation-plan.md b/docs/architecture/projection-distillation-plan.md index e0abc3f..a536f15 100644 --- a/docs/architecture/projection-distillation-plan.md +++ b/docs/architecture/projection-distillation-plan.md @@ -218,20 +218,6 @@ Minimum useful milestone: - deterministic input-slice builder - evaluator prompt or script that catches the known failure modes -## Near-Term Implementation Tasks - -1. Define `schemas/projection-example.schema.json`. Done. -2. Add `examples/projections/call-of-the-void.scene-broken-taxi-oz.jsonl` with - the first reviewed Cat and Oz examples. Done. -3. Add a lore grounding digest format for historical Aetheria projection - fixtures. Done. -4. Choose the first authored historical flashpoint for grounded training. -5. Add a deterministic projection input-slice tool that reads the agent-state - fixture and emits speaker-local input JSON. -6. Add a teacher prompt template for producing projection artifacts. -7. Add an evaluator prompt template for projection audits. -8. Only then consider student fine-tuning. - ## Boundaries The student projector is not the character. It is not the dialogue writer. It is diff --git a/docs/architecture/training-plan.md b/docs/architecture/training-plan.md index 8a3b2c2..df2ddc6 100644 --- a/docs/architecture/training-plan.md +++ b/docs/architecture/training-plan.md @@ -556,12 +556,12 @@ Purpose: choose and render what a character does next from local context. Current implementation: -- Qwen/local model calls through structured tools -- manual review and capture receipts -- no-thinking strict calls by default for tool compliance -- future gold responder data should use isolated generation workers or - sub-agents that receive only the projected local packet, not coordinator - context +- responder packet and output schemas with validators +- reviewed packet, capture, and mutation receipts +- archived local-model calls kept as plumbing tests and negative examples +- future gold responder data should use isolated generation workers, + sub-agents, or external model calls that receive only the projected local + packet, not coordinator context Inputs: @@ -693,7 +693,7 @@ Purpose: interpret an observable event from one participant's local perspective. Current implementation: - manual reviewed appraisals in mutation receipts -- Qwen appraiser experiments exposed invalid path drift +- archived local-model appraiser experiments exposed invalid path drift Inputs: diff --git a/docs/experiments/cold-wake-story-lab.md b/docs/experiments/cold-wake-story-lab.md index bc037ed..959045b 100644 --- a/docs/experiments/cold-wake-story-lab.md +++ b/docs/experiments/cold-wake-story-lab.md @@ -1,5 +1,16 @@ # Cold Wake Story Lab +Status: archived local-model/projection experiment. This doc records the early +Cold Wake story-lab run, including Qwen outputs, projection failures, and useful +pipeline lessons. It is not the current gold-data path. + +The current accepted-as-draft coordinator/IF scaffold is: + +- `examples/ink/cold-wake-branch-and-fold.v0.ink` +- `examples/ink/cold-wake-branch-and-fold.v0.training.json` +- `examples/coordinator/cold-wake-branch-and-fold.v0.json` +- `experiments/cold-wake-story-lab/cold-wake-branch-and-fold-clean-run.v0.md` + This experiment uses Ghostlight as a writing crutch on purpose. The goal is to write a small Cold Wake story while producing reusable projection @@ -72,11 +83,15 @@ preserves evidence and life without pretending the category system has healed. ## Experiment Loop +The archived run used this loop. Current gold-data runs should preserve exact +packet, capture, review, mutation, branch compiler, and IF reviewer seams rather +than treating model output as accepted prose by default. + For each scene: 1. Define character-local state and scene pressure. 2. Project the next response prompt for one character. -3. Send only `prompt_text` to the Qwen response model. +3. Send only `prompt_text` to the local response model. 4. Save the model output. 5. Review whether the output served the story, respected speaker-local context, and chose speech/action/silence plausibly. @@ -118,7 +133,7 @@ wording mistakes. The projection artifact now needs controls for frame ownership, authority boundaries, object custody, required compromise semantics, and forbidden resolutions before prompt prose is rendered. -## Qwen Target +## Archived Qwen Target Use the LAN Ollama endpoint from `gamecult-ops`: @@ -138,9 +153,9 @@ Use the LAN Ollama endpoint from `gamecult-ops`: ## Acceptance Criteria - The story has a visible arc, not just disconnected argument samples. -- Projection prompts include enough setting and cultural pressure for Qwen to - write without seeing the lore archive. +- Projection prompts include enough setting and cultural pressure for the + responder to write without seeing hidden author state. - At least one projected turn permits action instead of dialogue. -- Review notes explicitly mark whether Qwen chose speech, action, silence, or a - mixed beat. +- Review notes explicitly mark whether the responder chose speech, action, + silence, or a mixed beat. - Failures are kept when they teach a useful boundary. diff --git a/notes/documentation-inventory.md b/notes/documentation-inventory.md new file mode 100644 index 0000000..020df03 --- /dev/null +++ b/notes/documentation-inventory.md @@ -0,0 +1,93 @@ +# Documentation Inventory + +Use this file as the navigation map for Ghostlight's docs. It classifies files +by job so architecture, planning, archived receipts, raw imports, research +backlog, and state ledgers do not slowly melt into one cursed binder. + +## Canonical Re-Entry And Planning + +- `state/map.yaml`: canonical project map and current status. +- `notes/fresh-workspace-handoff.md`: re-entry summary for fresh sessions. +- `notes/ghostlight-current-system-map.md`: current runtime and artifact map. +- `notes/ghostlight-implementation-plan.md`: live implementation plan and + next-action surface. +- `notes/documentation-inventory.md`: this navigation/status index. + +## Core Architecture Contracts + +- `docs/architecture/canonical-agent-state-schema.md`: canonical state model. +- `docs/architecture/agent-state-variable-glossary.md`: meanings for state + axes, voice style, overlays, and emergent behaviors. +- `docs/architecture/agent-state-distributions-and-prompt-projection.md`: + deeper state-distribution and projection rationale. +- `docs/architecture/prompt-projection-contract.md`: projection boundary and + renderer contract. +- `docs/architecture/projected-local-context.md`: character-local context + shape. +- `docs/architecture/sequential-agent-runtime.md`: sequential action/reaction + runtime architecture. +- `docs/architecture/sandboxed-responder-packets.md`: isolated responder packet + contract and capture requirements. +- `docs/architecture/soft-model-training-artifacts.md`: rule for turning soft + judgments into training artifacts. +- `docs/architecture/ink-branching-scenes.md`: branch-and-fold, branch + compiler, visual prompt, and IF reviewer contract. +- `docs/architecture/lore-grounding-digest-format.md`: source-grounding digest + format. +- `docs/architecture/aetheria-lore-grounding-architecture.md`: Aetheria + historical, future-branch, and procedural grounding boundaries. +- `docs/architecture/technology-item-manifest-plan.md`: item, assembly, + supply-chain, trade-good, and blueprint data target. +- `docs/architecture/training-plan.md`: trainable organs, datasets, gates, and + rough scale targets. + +## Supporting Architecture And Future Mechanisms + +- `docs/architecture/aetheria-agent-population-model.md`: population and + faction-scale agent modeling. +- `docs/architecture/aetheria-content-generation-targets.md`: Aetheria content + targets and consumer use cases. +- `docs/architecture/aetheria-persistent-agent-architecture.md`: broader + persistent-agent architecture. +- `docs/architecture/projection-distillation-plan.md`: student projector + training concept. +- `docs/architecture/schema-future-mechanisms.md`: future schema organs waiting + for more evidence. +- `docs/architecture/aetheria-cold-wake-training-fixture.md`: Cold Wake fixture + notes and source-grounded room candidates. +- `notes/architecture-rationale.md`: rationale notes for major design + boundaries. + +## Archived Experiments And Plumbing + +- `docs/experiments/cold-wake-story-lab.md`: archived local-model/projection + experiment and failure lessons. +- `docs/architecture/qwen-invocation-notes.md`: archived local-model invocation + plumbing notes. + +## Raw Imported Source + +- `docs/aetheria/call-of-the-void-brainstorm.md`: raw imported brainstorming + document. Encoding scars are preserved; do not treat this as canonical prose. + +## Research Backlog + +- `docs/research/personality-model-roadmap.md`: personality-model research + roadmap. +- `docs/research/personality-signal-classifier.md`: personality signal + classifier research. +- `docs/research/things-to-steal.md`: research backlog and candidate ideas. + +## State Ledgers + +- `state/evidence.jsonl`: distilled current evidence ledger. +- `state/evidence.archive.jsonl`: archival evidence ledger. +- `state/branches.json`: branch ledger. +- `state/scratch.md`: disposable one-subgoal scratch surface. + +## Maintenance Rule + +Architecture docs describe durable contracts. Live planning belongs in +`notes/ghostlight-implementation-plan.md`, `notes/fresh-workspace-handoff.md`, +and `state/map.yaml`. Archived experiments can keep history, but they must be +labeled as archives when their path stops being the active route. diff --git a/notes/fresh-workspace-handoff.md b/notes/fresh-workspace-handoff.md index bd1f472..8b18b49 100644 --- a/notes/fresh-workspace-handoff.md +++ b/notes/fresh-workspace-handoff.md @@ -48,6 +48,7 @@ path is a source-grounded branching-scene data loop: - Canonical map: `state/map.yaml` - Current system map: `notes/ghostlight-current-system-map.md` - Implementation plan: `notes/ghostlight-implementation-plan.md` +- Documentation inventory: `notes/documentation-inventory.md` - Agent-state schema: `schemas/agent-state.schema.json` - Projected local context schema: `schemas/projected-local-context.schema.json` - Coordinator artifact schema: `schemas/coordinator-artifact.schema.json` diff --git a/notes/ghostlight-implementation-plan.md b/notes/ghostlight-implementation-plan.md index ba1a2b1..3892da4 100644 --- a/notes/ghostlight-implementation-plan.md +++ b/notes/ghostlight-implementation-plan.md @@ -6,6 +6,10 @@ Build the first reliable data-generation loop for socially persistent Aetheria agents. The immediate target is not a full simulation. It is a clean, reviewed, sandboxed training-data pipeline for branching scenes and state consequences. +Planning belongs in this file and the handoff/map state surfaces. Architecture +docs should describe durable contracts, boundaries, and data shapes; they should +not carry live next-action lists. + ## Near-Term Sequence 1. Stabilize canonical state and projection seams. @@ -66,8 +70,9 @@ sandboxed training-data pipeline for branching scenes and state consequences. 6. Train only after schemas stop sliding. - Pilot samples are schema shakedown, not robust corpus scale. - Specialized future models include coordinator, retriever, projector, - responder, appraiser, mutator, relationship/perception updater, evaluator, - and institution/faction/consumer decision models. + responder, appraiser, mutator, relationship/perception updater, branch + compiler, IF artifact reviewer, evaluator, and + institution/faction/consumer decision models. - Deterministic gates remain code: visibility, action legality, object custody, resource accounting, schema validation, source provenance, mutation authority, prompt leakage checks, and Ink compilation. diff --git a/state/evidence.jsonl b/state/evidence.jsonl index 7a44b4e..f3799a0 100644 --- a/state/evidence.jsonl +++ b/state/evidence.jsonl @@ -35,3 +35,4 @@ {"ts": "2026-05-02T22:29:00+00:00", "type": "design", "status": "ok", "note": "IF artifact reviewer required before acceptance to catch Potemkin state variables, cosmetic choices, fake folds, missing state reads, state-named-instead-of-checked prose, visual callback gaps, and endings that ignore major state."} {"ts": "2026-05-02T22:30:27+00:00", "type": "design", "status": "ok", "note": "Branch Compiler documented as a separate organ: it materializes playable Ink, sidecars, fold/callback structure, and visual handles from fixture state and reviewed consequence packets, without owning hidden character truth or durable social mutation."} {"ts": "2026-05-02T22:40:03+00:00", "type": "documentation", "status": "ok", "note": "Documentation cleanup aligned training plan, system map, handoff, soft-artifact doctrine, Ink architecture, Qwen notes, and content targets with current branch compiler + IF reviewer pipeline; old Qwen/local-model paths are archived plumbing, not active gold-data steering."} +{"ts": "2026-05-02T22:51:11+00:00", "type": "documentation", "status": "ok", "note": "Completed iterative documentation cleanup: removed live planning from architecture docs, archived local-model Cold Wake experiment wording, aligned training-plan responder/appraiser status, added documentation inventory, and restored next action to broader Aetheria fixture coverage."} diff --git a/state/map.yaml b/state/map.yaml index b13e08d..16cc73f 100644 --- a/state/map.yaml +++ b/state/map.yaml @@ -102,6 +102,7 @@ coverage_targets: current_artifacts: canonical_docs: + - notes/documentation-inventory.md - docs/architecture/prompt-projection-contract.md - docs/architecture/projected-local-context.md - docs/architecture/sequential-agent-runtime.md @@ -165,5 +166,5 @@ data_flow: - If state grows faster than understanding, cut it back. current_status: - summary: Cold Wake now has a clean accepted-as-draft branch-and-fold interactive-fiction scaffold. The run starts at the Ganymede route freeze tribunal, folds through Veyr provenance and Sella's clinic, carries compact consequence variables into the diagnostic-contact climax, and includes durable base image prompts plus branch/state modifiers. It is coordinator/IF scaffold training data, not raw responder gold. + summary: Cold Wake now has a clean accepted-as-draft branch-and-fold interactive-fiction scaffold, and the docs have a navigation inventory separating architecture contracts, planning surfaces, archived experiments, raw imported source, research backlog, and state ledgers. next_action: Move to a different Aetheria setting fixture. Pick the next faction, movement, flashpoint, or day-in-the-life target from the coverage plan, source-ground it in AetheriaLore, then build a similarly compact branch-and-fold IF scaffold with visual prompt anchors and reviewed consequence surfaces.