78 lines
3.2 KiB
Markdown
78 lines
3.2 KiB
Markdown
# HeroLib AI Prompts (`aiprompts/`)
|
|
|
|
This directory contains AI-oriented instructions and manuals for working with the Hero tool and the `herolib` codebase.
|
|
|
|
It is the **entry point for AI agents** that generate or modify code/docs in this repository.
|
|
|
|
## Scope
|
|
|
|
- **Global rules for AI and V/Hero usage**
|
|
See:
|
|
- `herolib_start_here.md`
|
|
- `vlang_herolib_core.md`
|
|
- **Herolib core modules**
|
|
See:
|
|
- `herolib_core/` (core HeroLib modules)
|
|
- `herolib_advanced/` (advanced topics)
|
|
- **Docusaurus & Site module (Hero docs)**
|
|
See:
|
|
- `docusaurus/docusaurus_ebook_manual.md`
|
|
- `lib/web/docusaurus/README.md` (authoritative module doc)
|
|
- `lib/web/site/ai_instructions.md` and `lib/web/site/readme.md`
|
|
- **HeroModels / HeroDB**
|
|
See:
|
|
- `ai_instructions_hero_models.md`
|
|
- `heromodel_instruct.md`
|
|
- **V language & web server docs** (upstream-style, mostly language-level)
|
|
See:
|
|
- `v_core/`, `v_advanced/`
|
|
- `v_veb_webserver/`
|
|
|
|
## Sources of Truth
|
|
|
|
For any domain, **code and module-level docs are authoritative**:
|
|
|
|
- Core install & usage: `herolib/README.md`, scripts under `scripts/`
|
|
- Site module: `lib/web/site/ai_instructions.md`, `lib/web/site/readme.md`
|
|
- Docusaurus module: `lib/web/docusaurus/README.md`, `lib/web/docusaurus/*.v`
|
|
- DocTree client: `lib/data/doctree/client/README.md`
|
|
- HeroModels: `lib/hero/heromodels/*.v` + tests
|
|
|
|
`aiprompts/` files **must not contradict** these. When in doubt, follow the code / module docs first and treat prompts as guidance.
|
|
|
|
## Directory Overview
|
|
|
|
- `herolib_start_here.md` / `vlang_herolib_core.md`
|
|
Global AI rules and V/Hero basics.
|
|
- `herolib_core/` & `herolib_advanced/`
|
|
Per-module instructions for core/advanced HeroLib features.
|
|
- `docusaurus/`
|
|
AI manual for building Hero docs/ebooks with the Docusaurus + Site + DocTree pipeline.
|
|
- `instructions/`
|
|
Active, higher-level instructions (e.g. HeroDB base filesystem).
|
|
- `instructions_archive/`
|
|
**Legacy / historical** prompt material. See `instructions_archive/README.md`.
|
|
- `todo/`
|
|
Meta design/refactor notes (not up-to-date instructions for normal usage).
|
|
- `v_core/`, `v_advanced/`, `v_veb_webserver/`
|
|
V language and web framework references used when generating V code.
|
|
- `bizmodel/`, `unpolly/`, `doctree/`, `documentor/`
|
|
Domain-specific or feature-specific instructions.
|
|
|
|
## How to Treat Legacy Material
|
|
|
|
- Content under `instructions_archive/` is **kept for reference** and may describe older flows (e.g. older documentation or prompt pipelines).
|
|
Do **not** use it as a primary source for new work unless explicitly requested.
|
|
- Some prompts mention **Doctree**; the current default docs pipeline uses **DocTree**. Doctree/`doctreeclient` is an alternative/legacy backend.
|
|
|
|
## Guidelines for AI Agents
|
|
|
|
- Always:
|
|
- Respect global rules in `herolib_start_here.md` and `vlang_herolib_core.md`.
|
|
- Prefer module docs under `lib/` when behavior or parameters differ.
|
|
- Avoid modifying generated files (e.g. `*_ .v` or other generated artifacts) as instructed.
|
|
- When instructions conflict, resolve as:
|
|
1. **Code & module docs in `lib/`**
|
|
2. **AI instructions in `aiprompts/`**
|
|
3. **Archived docs (`instructions_archive/`) only when explicitly needed**.
|