agentleFS
intent-driven-devintent-driven-dev/intent-driven-template/.agents/skills/architectural-decision-records/SKILL.md

architectural-decision-records

Skill185 starsChanged 32 days agoMITView on GitHub

Use when documenting, drafting, reviewing, or updating architectural decisions, ADRs, decision logs, tradeoffs, rationale, consequences, alternatives, or architecture decision history.

4 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

9 rules. 0 are common, 4 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use when creating, updating, reviewing, or explaining an ADR, architecture decision, decision log, tradeoff analysis, rationale, …

    line 14
  2. Distinctive:

    Do not use ADRs for transient implementation details, meeting notes, or insignificant decisions.

    line 16
  3. Very specific:

    Check preferences.md for preferred-style.

    line 21
  4. Very specific:

    If preferred-style: unset, ask the user which style to use, then record it in preferences.md before drafting.

    line 22
  5. Very specific:

    Use the matching file under templates/.

    line 23
  6. Very specific:

    If facts are missing, mark them as Unknown or ask a focused question.

    line 25
  7. Distinctive:

    Write honest consequences: benefits, downsides, follow-up.

    line 26
  8. Distinctive:

    If no preference is recorded, ask:

    line 51
  9. Very specific:

    After the user answers, update preferences.md with the selected preferred-style.

    line 55

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Architectural Decision Records

Overview

An Architectural Decision Record captures one architecturally significant decision, its rationale, tradeoffs, and consequences. Optimize for future readers reconstructing decision history.

When to Use

Use when creating, updating, reviewing, or explaining an ADR, architecture decision, decision log, tradeoff analysis, rationale, consequences, alternatives, or status.

Do not use ADRs for transient implementation details, meeting notes, or insignificant decisions. If significance is unclear, ask what future maintainers will need to know.

Workflow

  1. Identify the single decision. Split multiple decisions into multiple ADRs.
  2. Check preferences.md for preferred-style.
  3. If preferred-style: unset, ask the user which style to use, then record it in preferences.md before drafting.
  4. Use the matching file under templates/.
  5. Capture known facts only: context, requirements, constraints, options, rationale, decision makers, consequences.
  6. If facts are missing, mark them as Unknown or ask a focused question; do not invent context, options, or quality attributes.
  7. Write honest consequences: benefits, downsides, follow-up.
  8. Preserve history: supersede old accepted ADRs; do not rewrite them away.

Template Choice

SituationTemplate
Detailed tradeoff recordtemplates/madr-full.md
Concise tradeoff recordtemplates/madr-minimal.md
Classic lightweight ADRtemplates/nygard.md
One-sentence summarytemplates/y-statement.md
Project-specific styletemplates/custom.md
Too little evidenceADR review notes/questions, not accepted ADR

Status Values

StatusUse when
ProposedUnder review
AcceptedTeam committed
DeprecatedHistorically relevant but no longer recommended
SupersededReplaced by another ADR; link it

Preference Prompt

If no preference is recorded, ask:

Which ADR style should I use and remember for this project: MADR full, MADR minimal, Nygard, Y-Statement, or custom?

After the user answers, update preferences.md with the selected preferred-style. Do not ask again unless the user asks to change style.

Review Checklist

  • One decision, not a bundle
  • Significant: affects structure, quality attributes, constraints, or evolution
  • Context explains why it existed
  • Rejected options/tradeoffs are explicit
  • Rationale is tied to requirements, not preference
  • Downsides and follow-up work are recorded
  • Status is clear
  • Unknowns are marked, not invented

Example

# ADR-012: Use PostgreSQL for Orders

## Status

Accepted

## Context

Orders need relational constraints, consistency, and reporting joins. The team operates PostgreSQL well.

## Considered Options

- PostgreSQL: integrity, SQL reporting, familiar operations; migrations required.
- MongoDB: flexible schema; weaker fit for consistency and joins.

## Decision

We will use PostgreSQL because consistency, joins, and operational familiarity matter more than schema flexibility.

## Consequences

- Positive: Integrity and reporting align with needs.
- Negative: Schema changes need migrations.
- Follow-up: Define migration practice.

Common Mistakes

MistakeFix
Several decisions in one ADRSplit them.
Sales pitchInclude rejected options and negative consequences.
Invented contextMark unknowns or ask.
Treating status as decorationUse status to show lifecycle and link superseding ADRs.
Rewriting historyKeep old ADR; create/link superseding ADR.
Omitting alternatives under pressureInclude at least the rejected option and why it lost.

Sources

adr.github.io: home, ADR templates, AD practices. See templates/ for captured template variants.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.