108 lines
4.3 KiB
Markdown
108 lines
4.3 KiB
Markdown
# Decisions
|
|
|
|
status: draft_decision_log
|
|
date: 2026-06-20
|
|
|
|
## D001: New Repository Boundary
|
|
|
|
Decision: `Cognitive-OS-Wantsong` is a new repository and not a continuation branch of `the-mindscape-of-bro-tsong`.
|
|
|
|
Reason: the old project is useful as a governance lab and seed archive, but its process gravity is not suitable for proving the runtime product question.
|
|
|
|
## D002: Runtime-First Startup Rule
|
|
|
|
Decision: the project follows `First prove cognitive processing. Only then govern the pieces that prove useful.`
|
|
|
|
Reason: the old failure mode was proving controlled model governance while postponing the real cognitive-processing loop.
|
|
|
|
## D003: M0-M1 Only
|
|
|
|
Decision: this startup creates only project boundary and model-management kernel files.
|
|
|
|
Reason: sample runs, prompts, examples, and review packets belong to later milestones.
|
|
|
|
## D004: No Lite Filenames
|
|
|
|
Decision: model filenames and model IDs do not use a `lite` suffix.
|
|
|
|
Reason: the new project is not a light branch of a completed old runtime. Minimality is expressed as `runtime_scope: minimal_v0`.
|
|
|
|
## D005: QPI Role
|
|
|
|
Decision: QPI is a `routing_model` for problem framing, scarcity diagnosis, and misframing detection.
|
|
|
|
Reason: QPI must not become the main governance object again.
|
|
|
|
## D006: Intellectual Archaeology Role
|
|
|
|
Decision: Intellectual Archaeology is the first `depth_model` and may be the primary model for complex high-value issues.
|
|
|
|
Reason: the new runtime must test real deep processing instead of mostly guarding the entrance to deep processing.
|
|
|
|
## D007: Local CCRA Lane
|
|
|
|
Decision: Local CCRA remains a milestone review lane only.
|
|
|
|
Reason: review quality remains useful, but default round factories and Web bundle production are outside the startup scope.
|
|
|
|
## D008: No Old Process Artifact Migration
|
|
|
|
Decision: old `reports/`, `ccra_review_bundle/`, `local_ccra_reviews/`, full regression, full selector calibration, lifecycle scan workflows, and Round Conductor are not migrated.
|
|
|
|
Reason: those artifacts represent the old project gravity and would recreate the failure mode.
|
|
|
|
## D009: Registry Before Selector
|
|
|
|
Decision: `models/model-registry.json` records model metadata and orchestration hooks, but no selector is implemented in M0-M1.
|
|
|
|
Reason: the system needs model identity and call boundaries before it needs automation.
|
|
|
|
## D010: Reader Translation As A Core Layer
|
|
|
|
Decision: reader translation is a first-class product layer, even though its prompt is not created in M0-M1.
|
|
|
|
Reason: internal model output is not automatically understandable or usable by frontend readers.
|
|
|
|
## D011: M2 Prompt Authority Boundary
|
|
|
|
Decision: M2 must not directly implement GPT-authored prompt plans as canonical prompts in this repository.
|
|
|
|
Reason: canonical Prompt Cards, Agent Specs, Runtime Specs, model-call protocols, evaluation rubrics, and Local CCRA review protocol belong in `ccpe-system`.
|
|
|
|
## D012: Product-Local Adapter Scope
|
|
|
|
Decision: this repository may later hold only product-local draft adapters, sample runs, product workflow glue, and output formats.
|
|
|
|
Reason: product-local adapters are implementation-facing runtime surfaces, not canonical CCPE prompt authority.
|
|
|
|
## D013: Mandatory Adapter Frontmatter
|
|
|
|
Decision: if `prompts/` is created before CCPE supplies canonical contracts, every prompt-like file must declare:
|
|
|
|
```yaml
|
|
authority: product_local_adapter
|
|
status: draft
|
|
canonical_source: pending_ccpe
|
|
not_canonical_ccpe_prompt: true
|
|
```
|
|
|
|
Reason: future sessions must not confuse product-local adapters with canonical CCPE artifacts.
|
|
|
|
## D014: CCPE Request Channel For M2 Contracts
|
|
|
|
Decision: M2 canonical prompt/spec/rubric/protocol needs are recorded under `requirements/ccpe/`.
|
|
|
|
Reason: missing CCPE-owned authority should become a supplier request, not an ad hoc local implementation.
|
|
|
|
## D015: Skills-Vault Request Channel For Deterministic Tools
|
|
|
|
Decision: deterministic M2 tooling needs are recorded under `requirements/skills-vault/`.
|
|
|
|
Reason: reusable automation such as adapter metadata validation, run logging, and manifest checking belongs in `skills-vault`, not in this product repository.
|
|
|
|
## D016: Local CCRA Waits Until After A Minimal Manual Sample
|
|
|
|
Decision: Local CCRA must not start during M2 boundary correction.
|
|
|
|
Reason: Local CCRA should review milestone output after boundary correction and one minimal manual sample, not review the boundary correction itself as a new round.
|