agentleFS
broneqbroneq/bdk/skills/create-adr/SKILL.md

create-adr

Skill1 starsChanged 48 days agoMITView on GitHub

Create Architecture Decision Records (ADRs) based on MADR template. Use when the user asks to "create an ADR", "document a decision", "write an ADR", or provides decision context that needs to be formalized.

16 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 13 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use AskUserQuestion to fill gaps — only ask what's genuinely unclear.

    line 38
  2. Very specific:

    Write to docs/adr/NNNN-{slugified-title}.md following MADR format:

    line 52
  3. Very specific:

    Use ✅ for pros, ❌ for cons, 🟡 for neutral

    line 103
  4. Distinctive:

    Never use "Good, because" / "Bad, because" — use symbols directly

    line 104
  5. Distinctive:

    Only create diagrams when they add genuine visual value:

    line 108
  6. Very specific:

    Follow /bdk:mermaid-drawer for type selection, node budget and colour - an ADR outlives the person who wrote it, so its diagrams have to …

    line 116
  7. Distinctive:

    Read generated ADR file and verify formatting

    line 120

“Common” is measured: kept by at least 6 in 10 of the 13 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

Create ADR

Relies on BDK foundation (STARTUP_INSTRUCTIONS.md) for project context.

!python3 ${CLAUDE_PLUGIN_ROOT}/scripts/inject-rules.py architecture

Generate ADR following MADR format.

Input

$ARGUMENTS

Process

1. Analyze Input

Extract from free-form description:

  • Problem statement and context
  • Considered options
  • Constraints and decision drivers
  • Stated preference or chosen option

2. Ask Targeted Questions

Use AskUserQuestion to fill gaps — only ask what's genuinely unclear.

Always ask:

  • ADR status (proposed/accepted/rejected/deprecated)

Never ask — user adds manually after generation:

  • Decision-makers, consulted, informed (use {TBD} placeholder in frontmatter)

3. Determine Next ADR Number

Scan docs/adr/ for existing ADR files matching pattern NNNN-*.md. Use next sequential number, zero-padded to 4 digits.

4. Generate ADR Document

Write to docs/adr/NNNN-{slugified-title}.md following MADR format:

---
status: {status}
date: {YYYY-MM-DD}
decision-makers: {TBD}
consulted: {TBD}
informed: {TBD}
---

# ADR-NNNN: {Title}

## Context and Problem Statement
{2-3 sentences describing the problem}

## Decision Drivers
- {driver 1}
- {driver 2}

## Considered Options
1. **{Option A}** — {one-line summary}
2. **{Option B}** — {one-line summary}

## Decision Outcome
**Chosen option: {Option}**, because {justification}.

### Consequences
- ✅ {positive consequence}
- ❌ {negative consequence}
- 🟡 {neutral consequence}

### Implementation Requirements
- [ ] {requirement 1}

## Pros and Cons of the Options

### {Option A}
{description}

- ✅ {pro}
- ❌ {con}

### {Option B}
...

## More Information
{additional context, links, follow-up decisions}

Formatting rules:

  • Use ✅ for pros, ❌ for cons, 🟡 for neutral
  • Never use "Good, because" / "Bad, because" — use symbols directly

5. Generate Diagrams (Optional)

Only create diagrams when they add genuine visual value:

  • Architecture options with different component layouts
  • Options with different data flows or system interactions

Skip for trivially simple or abstract decisions.

If creating diagrams:

  • Embed Mermaid diagrams directly in ADR as fenced ```mermaid code blocks — renders natively, no compile step
  • Follow /bdk:mermaid-drawer for type selection, node budget and colour - an ADR outlives the person who wrote it, so its diagrams have to stay legible in whatever renderer a future reader opens them in

6. Final Verification

  • Read generated ADR file and verify formatting
  • Show user generated ADR path and summary

Rules

  • Diagrams optional — only generate when they add genuine visual value
  • Don't over-ask — extract maximum info from free-form input before asking
  • No commits — generate files only, user decides when to commit

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.