Creates, structures, and reviews technical documentation following the Diátaxis framework (tutorials, how-to guides, reference, and explanation pages). Use when a user needs to write or reorganize docs, structure a tutorial vs. a how-to guide, build reference docs or API documentation, create explanation pages, choose between Diátaxis documentation types, or improve existing documentation structure. Trigger terms include: documentation structure, Diátaxis, tutorials vs how-to guides, organize docs, user guide, reference docs, technical writing.
91
87%
Does it follow best practices?
Impact
96%
1.11xAverage score across 5 eval scenarios
Passed
No known issues
{
"context": "Tests whether the agent follows explanation documentation Diátaxis patterns: concept-framed title, Context→Core concept→Alternatives/trade-offs→Higher-level perspective structure, avoiding step-by-step instruction, and enabling the reader to understand the 'why' rather than just the 'what'.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Concept-framed title",
"description": "The document title is framed as a concept using patterns like 'How X works', 'Understanding X', or 'Why X is designed this way' — NOT as a task ('How to index') or a noun catalogue ('Index types')",
"max_score": 10
},
{
"name": "Context section present",
"description": "The document opens with a context section that situates the concept — explaining the problem space or the situation in which indexes matter, before diving into the concept itself",
"max_score": 10
},
{
"name": "Core concept explained",
"description": "The document includes a section explaining the core concept of how indexes work (e.g., B-tree structure, how the query planner uses them) at a conceptual level",
"max_score": 10
},
{
"name": "Trade-offs covered",
"description": "The document explicitly discusses at least two trade-offs or situations where indexes can hurt rather than help (e.g., write overhead, table bloat, low-cardinality columns)",
"max_score": 12
},
{
"name": "Alternatives discussed",
"description": "The document mentions or compares at least two alternative index types or approaches (e.g., B-tree vs. hash, partial indexes, composite indexes) in terms of when each is appropriate",
"max_score": 10
},
{
"name": "Higher-level perspective",
"description": "The document includes a section or closing discussion offering a higher-level perspective — e.g., how the query planner decides, when to add vs. remove indexes, or the mental model for reasoning about indexing",
"max_score": 10
},
{
"name": "No step-by-step instructions",
"description": "The document does NOT contain a numbered list of steps for creating or managing indexes as a primary content element (brief illustrative SQL snippets are acceptable; a tutorial walkthrough is not)",
"max_score": 10
},
{
"name": "No technical spec mixing",
"description": "The document does NOT include a comprehensive reference table of all index types with their parameters, defaults, and syntax",
"max_score": 8
},
{
"name": "Rationale emphasis",
"description": "The document emphasises the 'why' behind design decisions — at least two instances where the document explains the reason behind a behaviour or recommendation, not just the behaviour itself",
"max_score": 10
},
{
"name": "No how-to framing",
"description": "The document does NOT begin with 'This guide shows how to...' or frame its purpose as solving a specific user problem — it frames its purpose as building understanding",
"max_score": 10
}
]
}