Contract framework with markdown-native contracts utilizing fenced YAML blocks

This commit is contained in:
2026-05-03 22:51:13 +02:00
parent 3cfda33bc9
commit e3e13ee45a
36 changed files with 2877 additions and 13 deletions

View File

@@ -0,0 +1,13 @@
---
document_type: adr
---
# Weak ADR
## Context
This is short.
## Deprecated Approach
This section should not be here.

View File

@@ -0,0 +1,23 @@
---
document_type: adr
status: accepted
---
# Use Markdown Contracts
## Context
The problem is that plain heading counts do not explain whether content is
useful. Authors and agents need a contract that names the expected sections and
the job each section must do.
## Decision
We will use markdown-native document contracts with deterministic diagnostics as
the foundation for generation, validation, and later semantic assessment.
## Consequences
The tool can check author intent before generation or review work continues.
Future adapters can add form prefill and LLM rubrics without replacing the core
diagnostic model.

View File

@@ -0,0 +1,14 @@
---
document_type: business-letter
recipient_name: Ada Lovelace
---
# Incomplete Letter
## Greeting
Hello,
## Body
Thanks.

View File

@@ -0,0 +1,25 @@
---
document_type: business-letter
recipient_name: Ada Lovelace
sender_name: Markitect Team
---
# Follow-Up Letter
## Greeting
Dear Ada Lovelace,
## Body
Thank you for the thoughtful discussion about structured Markdown generation.
We reviewed the requirements and will send a concise proposal that separates
document contracts, field prefill, validation diagnostics, and optional semantic
assessment. This keeps the implementation practical while leaving room for
future automation.
## Closing
Kind regards,
Markitect Team

View File

@@ -0,0 +1,15 @@
---
document_type: concept-note
concept_id: contract-diagnostic-model
status: maybe
---
# Contract Diagnostic Model
## Definition
A vague note.
## Relationships
It relates to other things.

View File

@@ -0,0 +1,24 @@
---
document_type: concept-note
concept_id: contract-diagnostic-model
status: draft
---
# Contract Diagnostic Model
## Definition
A contract diagnostic model is the shared representation for validation,
assessment, and repair findings emitted by Markitect pipeline tools.
## Assertions
The central claim is that authors and agents need one diagnostic vocabulary
across structural checks, field checks, metric bands, and semantic assessments.
Evidence comes from the way legacy Markitect scattered related failures across
different subsystems.
## Relationships
The model relates to document contracts, form fields, section specifications,
and future LLM rubric adapters.

View File

@@ -0,0 +1,18 @@
---
document_type: prd-frs
product: Markitect Tool
---
# Thin PRD
## Problem
The document is too vague.
## Goals
The goals are listed without criteria.
## Implementation Plan
Build something immediately.

View File

@@ -0,0 +1,31 @@
---
document_type: prd-frs
product: Markitect Tool
owner: Platform Architecture
---
# Markitect Tool PRD And FRS
## Problem
Markdown pipelines often check document shape without knowing whether the
sections contain the content needed by authors, reviewers, and generation
agents.
## Goals
The product should make document contracts testable. Success metrics include
clear diagnostics, stable CLI behavior, and examples that show how contracts
apply to real document types.
## Functional Requirements
- Load Markdown contract files with fenced YAML contract blocks.
- Check required fields and section presence.
- Report metric bands and deterministic assertions.
- Produce machine-readable and human-readable diagnostics.
## Non-Goals
The first release does not execute provider-specific LLM calls or provide a UI
form renderer.

View File

@@ -0,0 +1,14 @@
---
document_type: workplan
status: blocked
---
# Weak Workplan
## Purpose
There is not enough implementation shape here.
## Tasks
The task list is prose only.

View File

@@ -0,0 +1,26 @@
---
document_type: workplan
id: MKTT-WP-EXAMPLE
status: active
---
# Example Workplan
## Purpose
Establish a focused implementation slice with enough structure for State Hub,
human review, and follow-on implementation.
## Tasks
```task
id: MKTT-WP-EXAMPLE-T001
status: todo
priority: high
```
Implement the smallest practical behavior and verify it through the CLI.
## Decision Point
Continue only if diagnostics are clear enough for humans and agents.