write-adr
Skill0 starsChanged 19 days agoMITView on GitHub
Write evidence-based architecture decision records, design documents, and RFC, using current code, decision history, realistic alternatives, documents, attachments, image, mermaid, diagram, infographic, brainstorm output, user discussion, and clear review conditions. Use when asked to draft, review, consult, accept, or supersede an ADR, or to write or revise a technical proposal, design document, or RFC. Not for implementing the design, an implementation plan (write-plan), choosing an architecture style (super-architecture), or documenting an existing API contract (create-documentation).
12 other projects use a version of this.
Adopted by
Most recently taken up by
- LeonardoEnnes/ai-engineering-toolkittheir own version, first seen Nov 8, 2026
- adhamhaithameid/write-skillstheir own version, first seen Nov 8, 2026
- helmut-hoffer-von-ankershoffen/heinzeltheir own version, first seen Nov 8, 2026
- 0xdeafcafe/skillstheir own version, first seen Nov 6, 2026
- daniel-kindl/dkstacktheir own version, first seen Nov 1, 2026
- tyabu12/claude-kittheir own version, first seen Oct 20, 2026
Other versions12 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 12
- HexamindOrganisation/hexgate★ 23same name
- helmut-hoffer-von-ankershoffen/heinzel★ 2same name
- adhamhaithameid/write-skills★ 1same name
- daniel-kindl/dkstack★ 1same name
- fabianmagrini/skills★ 1same name
- MaheshAwasare/claude-skills-pro★ 1same name
- tejastn10/workbench★ 1same name
- 0xdeafcafe/skills★ 0same name
- dr-zog/plumbline★ 0same name
- Industrial-AI-Research-Lab/nir-agent-overlay★ 0same name
2 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository57 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .cursor/rules/comments.mdcCursor rule
- .cursor/rules/commit-messages.mdcCursor rule
- .cursor/rules/git-workflow.mdcCursor rule
- .cursor/rules/manage-stack.mdcCursor rule
- .cursor/rules/no-worktrees.mdcCursor rule
- manage-stackSkill
- convert-as-skillSkill
- create-agentSkill
- create-agents-mdSkill
- create-commandSkill
- create-hookSkill
- create-ruleSkill
- create-skillSkill
- retrospectiveSkill
- super-architectureSkill
- super-cachingSkill
- super-elixirSkill
- super-nodejsSkill
- super-protobufSkill
- super-tech-blueprintSkill
- atomic-semantic-commitSkill
- code-reviewSkill
- create-bruno-collectionSkill
- create-documentationSkill
- create-ticketsSkill
- ready-specSkill
- write-changelogSkill
- write-mr-descriptionSkill
- write-planSkill
- write-postmortemSkill
- write-runbookSkill
- noslopSkill
- output-saversSkill
- safe-refactorSkill
- tddSkill
- verifySkill
- get-movingSkill
- later-listSkill
- standupSkill
- challenge-meSkill
- codebase-learnerSkill
- researchSkill
- second-opinionSkill
- traceSkill
- color-kitSkill
- design-system-kitSkill
- diagramSkill
- infographicSkill
- product-polish-workflowSkill
- prototype-wireframeSkill
- ui-ux-kitSkill
- evidenceSkill
- find-skillsSkill
- handoffSkill
- rumblingSkill
- use-levistackSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 12 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 47
Write a reviewable design proposal or record one consequential decision.
- Distinctive: line 64
Read the affected code, tests, plans, related ADRs, and available decision notes.
“Common” is measured: kept by at least 6 in 10 of the 12 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.
Files
Write an ADR or design document
Write a reviewable design proposal or record one consequential decision. Preserve the target repository's document format and accepted decision history.
Pick the job
| Request | Job |
| Draft or revise a broad design document, technical proposal, or RFC | Proposal: design-doc.md |
| Draft a formal architecture decision record | ADR draft: use the ADR procedure below |
| Check a proposed ADR | ADR review: use the ADR procedure below |
| Find the decision that governs a change | ADR consult: use the ADR procedure below |
| Accept, reject, deprecate, or replace an ADR | ADR lifecycle: lifecycle.md |
ADR review and consultation are read-only unless the user also asks for edits. A proposal does not become an accepted ADR by being well written.
ADR procedure
- Find existing ADRs, their index, filename or numbering scheme, status terms, and template. Follow that convention. If none exists, use the small fallback template in the existing docs tree without creating a fake initial decision or index.
- Read the affected code, tests, plans, related ADRs, and available decision notes. Inspect relevant Git history when it explains a prior choice. Separate current implementation from a proposal and an approved decision. Code presence alone does not prove acceptance.
- For a new record, check that an earlier ADR has not already made the decision. Cite or link the record when it has. Use
super-architecturewhen the architecture choice still needs evaluation; it returns the recommendation and evidence here without authoring the record. For a broader proposal, use this skill's Proposal job before recording an accepted choice.
Write or review the decision
State DARCI roles when known, context, and a distinct problem statement. Document each realistic approach in Approaches with a mermaid or diagram figure, pros, cons, and consequences; omit alternatives the evidence does not support. Name the decision in one sentence and tie it to the chosen alternative. Add reversibility or a revisit trigger when useful. Do not manufacture options merely to fill templates/adr.md.
Keep one decision per record. Label unknown rationale, missing authority, and open questions. Do not set accepted just because a draft is well written or implementation exists. When the repo requires a decision owner or date, use an evidenced value or leave a visible gap. Link the governing code and related records when available.
For Review, check the problem statement, DARCI gaps, decision, evidence, each approach (figure, pros, cons, consequences), status, links, and whether the record conflicts with another accepted decision. Report concrete gaps. Do not silently edit accepted history. For Consult, identify the governing accepted record and its scope; if records conflict or are superseded, show the chain and current uncertainty.
Verify the document
For an ADR, check the actual file and index for unique numbering, status consistency, links, source-backed claims, and bidirectional supersession links where the repo uses them. For a design proposal, check present-state evidence, acceptance conditions, unresolved decisions, and consistency between prose and visuals. A proposal with open questions is a valid draft, but it is not an accepted decision. Report what was checked and what remains unknown. Use evals/evals.json to exercise both jobs and routing boundaries.
Related skills
super-architecture: evaluate and choose a software architecture before recording it.write-plan: turn an accepted decision into implementation steps.create-documentation: document the implemented behavior that results from a decision.diagram: show an evidence-backed technical relationship or flow.infographic: show a sourced visual comparison or explanation while preserving current and proposed labels.polyglot-copywriter: refine the prose without changing technical facts.
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
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.

