Architecture Decision Record logo

Architecture Decision Record

Organization
Amazon-Quick
architecture-decision-record

Generates structured Architecture Decision Records (ADRs) from design conversations, meeting notes, or direct prompts. Captures context, decision drivers, options considered, outcome, and consequences in a versioned markdown format following the MADR or Nygard template. Use when asked to 'create an ADR', 'create architecture decision record', 'document this architecture decision', 'record why we chose X', 'write a decision record', or 'ADR for this design choice'.

Overview

PublisherAmazon-Quick
RepositoryAmazon-Quick-official-catalog
Skill namearchitecture-decision-record
Stars
50
Forks
4
Bundled files
1
LicenseMIT-0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by Amazon-Quick on GitHub. Read the source before you install it.

Installation

Install the Architecture Decision Record AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Amazon-Quick/Amazon-Quick-official-catalog/tree/main/skills/software-engineering/architecture-decision-record
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Amazon-Quick/Amazon-Quick-official-catalog.git /tmp/Amazon-Quick-official-catalog
mkdir -p .claude/skills
cp -r /tmp/Amazon-Quick-official-catalog/skills/software-engineering/architecture-decision-record .claude/skills/architecture-decision-record
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Architecture Decision Record in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Architecture Decision Record on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Architecture Decision Record is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Overview

Generates Architecture Decision Records from design conversations, meeting notes, or direct prompts. Walks through decision drivers, options considered, and consequences, then produces a numbered, versioned markdown file ready for commit to a docs repository.

Workflow

<Definition - ADR Status> Each ADR carries exactly one status at any time:

  • Proposed: Decision documented but not yet ratified by the team.
  • Accepted: Decision ratified and in effect.
  • Deprecated: Decision no longer applies due to changed circumstances but remains in the record for history.
  • Superseded: Decision replaced by a newer ADR. The superseded record links forward to its replacement, and the replacement links back. </Definition - ADR Status>

<Definition - MADR Format> Markdown Any Decision Record (MADR) is a lean, structured template popularized by the adr-tools community. Sections: Title, Status, Context and Problem Statement, Decision Drivers, Considered Options, Decision Outcome (with Rationale), Consequences (Positive, Negative, Neutral), and optional Links to related ADRs. </Definition - MADR Format>

<Definition - Nygard Format> Michael Nygard's original ADR format. Sections: Title, Status, Context, Decision, Consequences. Shorter and less prescriptive than MADR. Best for teams that prefer brevity over exhaustive structure. </Definition - Nygard Format>

<Definition - ADR Numbering> ADRs are numbered sequentially with zero-padded four-digit prefixes (e.g., 0001, 0002). The number is determined by scanning the target directory for existing ADR files and incrementing the highest found number by one. If no directory is specified or no existing ADRs are found, numbering starts at 0001. </Definition - ADR Numbering>

<Workflow - Generate ADR description="End-to-end ADR generation from input to written file." tools=[file_write, file_read, run_python, open_in_session_tab] triggers=["create an ADR", "document this architecture decision", "record why we chose X", "write a decision record", "ADR for this design choice"]

  1. [Ask user] Gather the decision title if not already provided. Confirm the template choice (madr, nygard, or custom). If context was supplied (transcript, notes, or description), acknowledge it. If no context was given, ask the user to describe the decision background in a few sentences. If fails: If the decision title or template choice is missing, re-ask the user for the missing item before continuing.

  2. [Decide] If the user selected "custom" template, ask them to describe or paste their preferred format. Store the structure for use in the drafting step. Otherwise, proceed with the selected standard template.

  3. [Think] Analyze the provided context. Extract: the core problem or question, factors that drove the decision (performance, cost, team expertise, timeline), distinct options that were discussed, the chosen option, and any stated trade-offs or risks. If the context is a long transcript, identify only the segments relevant to this specific decision.

  4. [Ask user] Present the extracted elements in a structured summary:

    • Problem Statement: (one or two sentences)
    • Decision Drivers: (bulleted list)
    • Options Considered: (numbered list with one-line descriptions)
    • Chosen Option: (name)
    • Key Rationale: (why this over the others)

    Ask the user to confirm, correct, or add missing items. Per Rule 1, if options are unclear, explicitly ask "Were there other alternatives discussed?" If fails: If the user does not confirm or correct the elements, re-present the summary and ask specifically which items to add or fix.

  5. [Ask user] Ask about consequences. Per Rule 6, prompt the user: "What positive outcomes do you expect from this decision? What are the downsides or risks?" If the user provides at least one of each, proceed. If not, suggest plausible consequences based on the context and ask for confirmation. If fails: If no consequences are provided or confirmed, propose at least one positive and one negative consequence from the context and ask the user to confirm.

  6. [Ask user] Confirm the status for this ADR. Default is "Proposed." Ask: "Is this decision already accepted by the team, or should it remain as Proposed for review?" If fails: If the user does not specify a status, default to Proposed and note that the status can be updated later.

  7. [Ask user] Ask where the ADR file should be saved. Offer to scan the workspace for existing ADR directories. If the user names a path, use it directly. If fails: If no save location is given, offer to scan the workspace for ADR directories and ask the user to pick one.

  8. [Agent] Scan the target directory for existing ADR files matching the pattern NNNN-*.md. Determine the next sequential number. If the directory does not exist, create it and start at 0001. If fails: If the directory cannot be scanned or created, report the path error to the user and ask for a valid location.

  9. [Decide] Does this ADR supersede an existing one? If the user indicated it replaces a prior decision, identify the old ADR file. Read its current status. Prepare to update the old file's status to "Superseded" with a forward link per Rule 7.

  10. [Think] Compose the full ADR document using the selected template. Populate every section with user-confirmed content. Use the appropriate template from the Templates block below. Ensure the file name follows Rule 8.

  11. [Ask user] Present the complete draft ADR in a code block. Ask: "Does this look correct? Any changes before I write the file?" Wait for explicit approval. Validate: User approves or requests edits. If edits requested, incorporate and re-present. If fails: If the user does not approve or request edits, re-present the draft and ask for explicit approval before writing.

  12. [Agent] Write the ADR file to the target directory with the computed file name. If this ADR supersedes another, update the old file's status line and append the "Superseded by" link. If fails: If the ADR file cannot be written, report the write error, retry once, and present the draft inline if the retry fails.

  13. [Agent] Open the newly created ADR in the session tab for the user to review in rendered markdown. If fails: If the ADR cannot be opened in the session tab, report the error and provide the file path so the user can open it manually.

  14. [Agent] Present a summary: file path, ADR number, status, and a reminder to check for numbering collisions before merging (per Gotchas). If fails: If the summary cannot be presented, report the error and restate the file path and ADR number directly.

</Workflow - Generate ADR>

<Template - MADR>

markdown
# {{number}}. {{decision_title}}

Date: {{date}}

## Status

{{status}}

## Context and Problem Statement

{{context_and_problem}}

## Decision Drivers

{{decision_drivers_bulleted}}

## Considered Options

{{options_numbered}}

## Decision Outcome

Chosen option: "{{chosen_option}}", because {{rationale}}.

### Positive Consequences

{{positive_consequences_bulleted}}

### Negative Consequences

{{negative_consequences_bulleted}}

## Links

{{links_to_related_adrs}}

</Template - MADR>

<Template - Nygard>

markdown
# {{number}}. {{decision_title}}

Date: {{date}}

## Status

{{status}}

## Context

{{context_paragraph}}

## Decision

{{decision_paragraph}}

## Consequences

{{consequences_paragraph}}

</Template - Nygard>

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Architecture Decision Record AI skill do?

Generates structured Architecture Decision Records (ADRs) from design conversations, meeting notes, or direct prompts. Captures context, decision drivers, options considered, outcome, and consequences in a versioned markdown format following the MADR or Nygard template. Use when asked to 'create an ADR', 'create architecture decision record', 'document this architecture decision', 'record why we chose X', 'write a decision record', or 'ADR for this design choice'.

Why use Architecture Decision Record on TypingMind?

Because you install it once and use it with any model. Architecture Decision Record is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Architecture Decision Record in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Amazon-Quick/Amazon-Quick-official-catalog/tree/main/skills/software-engineering/architecture-decision-record. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Architecture Decision Record?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Architecture Decision Record?

As many as you like. As long as a model supports skills, you can use Architecture Decision Record with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Architecture Decision Record AI skill free?

Yes. It is published on GitHub by Amazon-Quick under the MIT-0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇