Agent skill
imrad-recomposer
Use when transforming non-IMRaD material into a clean IMRaD output with strict section boundaries and explicit assumption labeling.
Install this agent skill to your Project
npx add-skill https://github.com/narumiruna/agent-skills/tree/main/skills/imrad-recomposer
SKILL.md
IMRaD Recomposer
Purpose
Transform non-IMRaD material into IMRaD-structured output.
When to use
Use this skill when the user provides:
- Notes, transcripts, explanations, or fragmented analysis
- Mixed reasoning that needs separation into IMRaD sections
- Research-style analytical content requiring a production-ready IMRaD draft
When NOT to use
Do NOT use this skill when:
- The detector recommendation is "do not use IMRaD"
- The user requests only diagnosis or review
- The task is simple Q&A, tutorials, or ideation without meaningful analytical structure
Input Assumptions
- Input may be partial or noisy
- Source evidence may be incomplete
- Assumptions are allowed only when explicitly stated
Core Workflow
- Extract the core question/problem.
- Derive or identify method/approach.
- Separate findings from interpretation.
- Compose sections in exact IMRaD order.
- Validate claim traceability and section boundaries.
Handling Incomplete Inputs
If methods or results are missing, the agent MUST:
- Infer a reasonable method only for analytical or design-oriented tasks
- State assumptions explicitly in Methods
- Use non-empirical results only when empirical results are unavailable
- Label non-empirical results as inferred, hypothetical, expected, or simulated
Truthfulness Constraints
- The agent MUST NOT fabricate empirical data, citations, experiments, measurements, or observations
- Every major claim in Results and Discussion MUST be traceable to explicit input evidence or explicit assumptions
Output Requirements
The output MUST:
- Follow IMRaD order exactly: Introduction, Methods, Results, Discussion
- Use clear section headers
- Keep boundaries clean between reporting and interpretation
- Be concise, precise, and non-redundant
- Prefer bullets where helpful
Section Specifications
Introduction
- Define context and question
- State objective or hypothesis
- MUST NOT include methods, results, or conclusions
Methods
- Describe approach, process, tools, and assumptions
- Include reproducibility details when available
- MUST NOT include interpretation or findings
Results
- Report findings objectively
- Separate empirical from non-empirical outputs
- MUST NOT include interpretation, implication, or argument
Discussion
- Interpret only what Results support
- Answer the question from Introduction
- Include limitations and concrete next steps when relevant
Failure Modes
- Mixing Results with Discussion
- Introducing claims unsupported by evidence or assumptions
- Hiding assumptions used to infer missing components
- Producing generic prose instead of structured analytical transformation
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
slide-creator
Use when creating slide decks with Marp/Marpit Markdown (marp), including authoring slide content, designing slide color schemes, and building SVG diagrams or illustrations for the deck.
slide-color-design
Use when you only need slide color systems, palette selection, or palette generation workflows and want direct pointers to the color design references.
mermaid-creator
Use when creating or converting Mermaid diagrams (for example flowcharts, sequence diagrams, ER diagrams, and Gantt charts), including exporting to SVG for docs or slides.
test-driven-development
Use when implementing non-trivial code changes that should follow TDD (write a failing test first, make the smallest passing change, then refactor safely).
svg-illustration
Use when you need SVG diagram rules, layout patterns, or embedding guidance for slide decks and want the minimal SVG-focused reading path.
python-cli-typer
Use when building or structuring Python CLI commands with Typer, including commands, options, and multi-command apps.
Didn't find tool you were looking for?