Architecture Decision logo

Architecture Decision

Community
davekilleen
architecture-decision

Use when a system or design choice needs an evidence-backed ADR, explicit alternatives and trade-offs, or review of proposed, accepted, or superseded decision history.

Overview

Publisherdavekilleen
RepositoryDex
Skill namearchitecture-decision
Stars
481
Forks
130
Bundled files
Instructions only
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.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by davekilleen on GitHub. Read the source before you install it.

Installation

Install the Architecture Decision 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.
  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/davekilleen/Dex.git /tmp/Dex
mkdir -p .claude/skills
cp -r /tmp/Dex/packages/dex-agent-plugin/skills/_available/engineering/architecture-decision .claude/skills/architecture-decision
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Architecture Decision 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 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 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.

Architecture decision

When to use

Use this skill when a material architecture choice, constraint, or reversal needs a traceable Architecture Decision Record (ADR), or when an existing ADR's state and rationale need review.

Do not use it for live incident response, implementation instructions that do not require a design decision, or a decision that has no identifiable evidence or human decision authority. Not for silently choosing on behalf of the team or rewriting historical ADRs.

Inputs and source discipline

Start with the decision question, scope, constraints, affected systems, decision authority, and intended decision window. Treat a missing authority, constraint, or date as unknown; ask for it or mark it TBD.

Build a source ledger before comparing options. For every material claim record the source, source date, as-of date/time, locator, and freshness or access limitation. Prefer directly inspected code, configuration, tests, measured operations, and dated requirements; label meeting recollections and interpretations accordingly. Keep conflicting sources linked rather than silently choosing one.

Method

  1. Read the relevant context in read-only mode. State the problem, desired outcome, non-negotiable constraints, decision deadline if evidenced, and what is outside scope.
  2. Enumerate the status quo and credible alternatives. For each option capture supporting evidence, fit to constraints, benefits, costs, risks, reversibility, migration or operational burden, trade-offs, and unresolved unknowns.
  3. Compare options with the same criteria. Record confidence for each material assessment and retain contradictions; do not manufacture a score, percentage, or cost when the source does not provide one.
  4. Draft the ADR with one explicit state:
    • Proposed means the evidence-backed draft is awaiting human authority.
    • Accepted means the human decision authority explicitly approved the stated option and trade-offs; a recommendation is not approval.
    • Superseded means a later accepted ADR replaces it, with links in both directions. Preserve the earlier record as immutable history.
  5. Present a recommendation as an option for the human decision authority. Do not change a proposed ADR to accepted, or create the canonical ADR, until that human authority confirms the action and the exact decision text.

Truth and uncertainty rules

Label each claim as observed, inferred, unknown, stale, or contradictory. Observed means the cited source directly supports it; inferred means reasoning from cited facts; unknown means the evidence is absent; stale means the source may no longer represent the decision context; contradictory means credible sources disagree.

Never invent dates, metrics, owners, intent, money, percentages, causes, status, or evidence. Do not turn an inference into a constraint or an unknown into a risk score. State which source would resolve each material unknown and do not hide disagreement behind a single confidence number.

Output contract

Return a decision brief or ADR containing:

  • decision question, scope, constraints, as-of date/time, and decision authority;
  • a source ledger with source, source date, citations, and freshness;
  • alternatives, consistent comparison criteria, evidence-backed trade-offs, and confidence;
  • state as Proposed, Accepted, or Superseded, with the approval or successor link required by that state;
  • unknowns and contradictions, including their effect on the recommendation;
  • implementation consequences and follow-up evidence, if known; and
  • a clearly labelled recommendation that is not a human decision.

Safety and write boundaries

The default is read-only. A write requires a preview of exact destination and content, then explicit confirmation from the human decision authority before creation or modification. Do not edit source code, configuration, tickets, or an accepted or superseded ADR as part of this review. Do not claim that a proposed option is adopted or send implementation instructions as if they were approval.

Keep ADR history append-only. If a historical record is wrong, preserve it and propose a correction or successor for human review; never erase the evidence of the earlier state. Recommendations are not human decisions, and no external action is authorized by this skill alone.

Verification and recovery

Before writing, read back the preview and reconcile the state, decision text, authority, source citations, option links, and destination. After a confirmed write, read back the persisted ADR and reconcile it with the approved preview and its successor or predecessor links. Check that no unsupported date, status, or evidence appeared.

If a read, write, or reconciliation check fails, stop and report the exact failure, partial state, and missing evidence. Do not retry blindly or overwrite history. Recover by re-reading the destination and source ledger; only a human-authorized append, correction, or superseding ADR may resolve a partial or contradictory record.

Frequently asked questions

What does the Architecture Decision AI skill do?

Use when a system or design choice needs an evidence-backed ADR, explicit alternatives and trade-offs, or review of proposed, accepted, or superseded decision history.

Why use Architecture Decision on TypingMind?

Because you install it once and use it with any model. Architecture Decision 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 in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/davekilleen/Dex/tree/main/packages/dex-agent-plugin/skills/_available/engineering/architecture-decision. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Architecture Decision?

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?

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

Is the Architecture Decision AI skill free?

It is published on GitHub by davekilleen. Check the repository for licensing terms. 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 👇