98 lines
3.3 KiB
Markdown
98 lines
3.3 KiB
Markdown
# Knowledge Asset Rules
|
|
|
|
## Purpose
|
|
|
|
`knowledge_assets/` stores stable, long-term, reusable knowledge documents for this repository.
|
|
|
|
These documents are not raw source archives, not temporary handoffs, and not implementation-only files. They are the durable explanation layer that helps the project owner, ChatGPT, Codex, and future CCRA work understand the model library without reconstructing context from scattered files.
|
|
|
|
For the broader repository-wide distinction between canonical, generated, review archive, and temporary files, see `docs/FILE_TAXONOMY.md`.
|
|
|
|
## What Belongs In `knowledge_assets/`
|
|
|
|
Use `knowledge_assets/` for:
|
|
|
|
- product context summaries
|
|
- model maps
|
|
- model card structure rules
|
|
- model extraction templates
|
|
- stability rating rules
|
|
- durable process records
|
|
- reusable architecture summaries
|
|
- durable data-governance and model-invocation mechanism explanations
|
|
|
|
Do not use `knowledge_assets/` for:
|
|
|
|
- raw original articles
|
|
- temporary validation reports
|
|
- full copies of every model card
|
|
- generated cache files
|
|
- implementation scripts
|
|
- source-of-truth JSON model assets
|
|
- per-round review bundles or optional raw changed-file zips
|
|
|
|
## Relationship To Other Directories
|
|
|
|
`docs/` contains operative local rules, contracts, workflow docs, and decisions.
|
|
|
|
`models/` contains machine-readable model JSON source of truth.
|
|
|
|
`cards/` contains human-readable model cards.
|
|
|
|
`reports/` contains session-level validation, audit, and handoff documents.
|
|
|
|
`knowledge_assets/` contains stable explanatory knowledge distilled from those sources.
|
|
|
|
The project owner may manually sync stable `knowledge_assets/` documents into GPT knowledge storage. Review bundles should therefore not include `knowledge_assets/` by default unless the current review explicitly targets the long-term knowledge asset itself.
|
|
|
|
## Naming Rule
|
|
|
|
Use numeric prefixes for reading order.
|
|
|
|
Do not put version numbers in filenames. Put version, status, and last updated date inside the document.
|
|
|
|
Examples:
|
|
|
|
```text
|
|
00_用户背景与产品上下文.md
|
|
01_核心模型地图.md
|
|
02_模型卡结构规范.md
|
|
03_核心模型抽取样板.md
|
|
06_模型稳固性评级规则.md
|
|
07_产品规划过程记录.md
|
|
09_数据治理与模型调用机制说明.md
|
|
```
|
|
|
|
## Model Card Sample Rule
|
|
|
|
Do not duplicate every concrete model card into `knowledge_assets/`.
|
|
|
|
Concrete model card source of truth remains in `cards/` and `models/`.
|
|
|
|
If a sample is needed for ChatGPT or CCRA reasoning, create a clearly marked sample or summary document, not a second source of truth.
|
|
|
|
For now, QPI and 思想考古 are referenced through:
|
|
|
|
- `cards/qpi.md`
|
|
- `cards/intellectual_archaeology.md`
|
|
- `models/qpi.model.json`
|
|
- `models/intellectual_archaeology.model.json`
|
|
|
|
## Update Rule
|
|
|
|
Update or create a knowledge asset when:
|
|
|
|
- a rule becomes stable enough to reuse across sessions
|
|
- project context changes in a durable way
|
|
- a model map changes
|
|
- a workflow or schema decision should be remembered outside a single report
|
|
- ChatGPT handoff material repeats across sessions
|
|
|
|
Do not update knowledge assets for every small implementation change. Use reports for session-level details.
|
|
|
|
## Review Rule
|
|
|
|
Knowledge assets should be concise and source-linked.
|
|
|
|
When a knowledge asset summarizes a file-backed artifact, link to the artifact and state whether the source of truth remains elsewhere.
|