Skip to content
Back to skills

Adr

ASecurity

Read and write Architecture Decision Records in docs/adr/. Use for /adr, when decisions are made, and before changing areas covered by ADRs.

  • 7 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 13, 2026
ai-agents

Works with

  • cli

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned November 13, 2026

npx -y skills add zazencodes/skills --skill adr --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Adr?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Adr
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/zazencodes-adr/badge)](https://www.skillsdirectory.com/skills/zazencodes-adr)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: adr
description: Read and write Architecture Decision Records in docs/adr/. Use for /adr, when decisions are made, and before changing areas covered by ADRs.
---

# ADR

An ADR records a standing decision so the next session neither undoes it nor argues it again. The code is the source of truth for how things work; an ADR keeps only what was decided, why, and what was ruled out.

## Where they live

- `docs/adr/`, one file per decision, named `YYYY-MM-DD-<topic>.md`. The topic is a short claim: `the-clip-title-names-the-clip`, `postgres-full-text-search`.
- A repo of independent projects keeps them per project in `<project>/docs/adr/`. Follow the repo's `AGENTS.md`.
- The project's `AGENTS.md` indexes them under `## Decisions`: one line per ADR, a short summary and the path.
- Create `docs/adr/` when writing the first one.

## Before changing an area

Read the index in `AGENTS.md`, open every ADR that touches the area, and work within them. If the task conflicts with one, stop and tell the user which ADR and how before writing code. Never work around an ADR quietly.

## When to write one

Write one whenever a real decision is made: a choice between viable approaches, a rejected alternative, or a constraint the next session would otherwise undo. Write it in the same change as the code. Skip choices with one obvious answer.

Invoked as `/adr` with nothing else, record the decision just made in the conversation. If none is clear, ask which one.

## Format

```md
# The decision, stated as a short claim

## Decision

What was decided, in a few sentences or bullets.

## Why

The reasons: the needs and constraints that settled it.

## Rejected

- **The alternative.** Why it lost.
```

Say what must not be re-litigated in Decision or Rejected. Keep it to one screen.

## Rules

- Stay high-level. Never restate function names, schemas, config values, geometry or anything else a refactor would change. If a refactor could make a sentence false while the decision still stands, cut the sentence.
- Supersede an ADR by editing or deleting it, never by writing a new one that contradicts it.
- Add, update or remove its line in the `AGENTS.md` index in the same change.

Files in this skill

  • SKILL.md2.1 KB
  • agents/openai.yaml150 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.