agentleFS
vinayagrwvinayagrw/claude-skills/architecture-decision-record/SKILL.md

architecture-decision-record

Skill1 starsChanged 25 days agoApache-2.0View on GitHub

Draft an ADP Architecture Decision Record (ADR) in Tango's official template, capturing a decision's context, drivers, considered options, trade-off, outcome, consequences, confirmation, and risks. Use when the user records a technical decision for the Agent Development Platform / Project Jedi, asks to "write an ADR", or needs a Confluence-paste-ready ADR for the ADP space.

13 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

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 10 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Only write one for decisions that are expensive to reverse, cross team/plane boundaries, or constrain future decisions.

    line 23
  2. Very specific:

    Fill every heading: Write N/A rather than deleting a heading — the shape stays consistent across records.

    line 29
  3. Distinctive:

    When the user wants this pasted into the ADP space, hand off to the

    line 62

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

ADP Architecture Decision Record (ADR)

Produces an ADR that matches Tango's official ADP ADR template (Confluence source). The full fill-in skeleton is in references/adp-adr-template.md — follow it heading-for-heading.

When to use this

  • Recording a decision for the ADP / Project Jedi platform.
  • The user says "write an ADR", "record this decision", or needs it in the ADP space.

When an ADR is warranted (gate first)

Only write one for decisions that are expensive to reverse, cross team/plane boundaries, or constrain future decisions. Anything smaller is a one-line note, not an ADR. If it doesn't clear that bar, say so instead of producing a record.

Rules that define this template

  1. One decision per record. An ADR captures why, not a design spec.
  2. Fill every heading. Write N/A rather than deleting a heading — the shape stays consistent across records.
  3. Immutability. Once Accepted, never rewrite the substance. A material change means a new ADR that supersedes this one; the supersession chain is the audit trail.
  4. No straw-man options. List only real alternatives that were genuinely on the table (2–5).
  5. Ratified date stays blank until a human signs off.

Structure (follow the reference skeleton exactly)

  1. Header table — Status, Draft date, Ratified date, Deciders, Affected planes, Affected capabilities, Review-by.
  2. Context and Problem Statement — 2–4 neutral sentences; phrase the core as a question ("How should ADP … ?"); no answer yet. Bullets: Requirement, Constraint, Non-functional target, Existing context.
  3. Decision Drivers — the criteria options are judged against (why it's hard).
  4. Considered Options — named list (Option A/B/C…).
  5. Per option — an assessment table (Complexity, Effort/cost, Portability, Vendor- neutrality, Reversibility, Persona impact), Architecture, Sequence Diagram, Pros, Cons. Mark the recommended one — RECOMMENDED.
  6. Trade-off Analysis — the decisive trade-off, honestly: what you give up to get what you want.
  7. Decision Outcome — chosen option + the decisive reason, tied to the drivers.
  8. Consequences — Good / Bad / Neutral-or-follow-on.
  9. Confirmation — how the decision will be verified (ideally machine-checkable): table of How we'll confirm / Owner / When.
  10. Risks & Technical Debt — table: Risk-or-debt / Likelihood / Impact / Mitigation.
  11. Action Items — ordered, each with owner + target date; always end with "Update the ADR Register with the final status".
  12. More Information — Prior art / SOTA, Related ADRs & Jira (JEDI-XXXX), Open questions deferred.

Reference values

  • Six planes: Build & Foundry · Data · Run & Channels · Defend & Trust · Govern & Deploy · Operate & Workforce.
  • Status lozenges: Proposed (blue) · Accepted (green) · Rejected (red) · Deprecated (yellow) · Superseded (grey/neutral).

Publishing to Confluence

When the user wants this pasted into the ADP space, hand off to the tango-confluence-ops skill for the actual page edit. Then honor those conventions: render Status cells as real status lozenges, dates as native date elements, people as real mentions, and use no em dashes.

Output

A complete ADR with every heading filled (or N/A), all tables populated, and at least two real options assessed — ready to copy into a new page under the ADP ADR Register.

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.