adr-generator
Skill11 starsChanged 49 days agoApache-2.0View on GitHub
Produces a complete inline ADR (Context, Decision, Status, Alternatives, Consequences). Use when a significant design decision is identified during design doc creation, when an engineer asks to document a decision as an ADR, or when an existing ADR is incomplete (missing alternatives, consequences, or status). Trigger phrases — "generate an ADR for this decision", "document this as an architecture decision record", "create an ADR for choosing X over Y".
7 other projects use a version of this.
Adopted by
Most recently taken up by
- nek1987/auto-agent-harnesstheir own version, first seen Oct 11, 2026
- patonkikh/APEStheir own version, first seen Jul 3, 2026
- hailcpy/gen-adrtheir own version, first seen Jun 9, 2026
- riosee2415/WITIVE-RAG-AGENTtheir own version, first seen May 8, 2026
- AntPerez69367/claude-skillstheir own version, first seen Jan 28, 2026
- romain325/skillstheir own version, first seen Jan 14, 2026
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 7
- hailcpy/gen-adr★ 6same name
- nek1987/auto-agent-harness★ 4same name
- patonkikh/APES★ 3same name
- abdullah1854/ClaudeSuperSkills★ 1same name
- romain325/skills★ 1same name
- AntPerez69367/claude-skills★ 0same name
- riosee2415/WITIVE-RAG-AGENT★ 0same name
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 repository83 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- about-konductorSkill
- adversarial-code-review-pass-integritySkill
- adversarial-code-review-pass-schemaSkill
- adversarial-code-review-pass-securitySkill
- adversarial-code-reviewSkill
- adversarial-design-reviewSkill
- agents-md-authoringSkill
- app-discoverySkill
- architecture-advisorSkill
- architecture-diagram-generationSkill
- argumentation-referenceSkill
- asana-sprint-planningSkill
- asdlc-aspect-reviewSkill
- asdlc-code-simplifierSkill
- aws-service-validatorSkill
- backend-developmentSkill
- backend-reviewSkill
- claude-teams-behaviorSkill
- cmux-dispatchSkill
- code-reviewSkill
- constraintsSkill
- cost-estimationSkill
- cypress-test-implementationSkill
- decision-researchSkill
- decision-writingSkill
- delegation-protocolSkill
- deliberation-panelSkill
- design-doc-guidelinesSkill
- design-evaluationSkill
- design-impact-reviewSkill
- design-quality-checkSkill
- doc-accuracy-analyzerSkill
- document-formatsSkill
- dom-inspectionSkill
- dynamodb-designSkill
- dynamodb-validationSkill
- e2e-test-strategySkill
- external-researchSkill
- find-aws-skillsSkill
- frontend-developmentSkill
- frontend-reviewSkill
- git-mergeSkill
- git-workflowSkill
- historical-issues-registrySkill
- humanize-writingSkill
- iam-policy-designSkill
- infra-validationSkill
- kiro-design-generationSkill
- kiro-requirements-generationSkill
- kiro-task-generationSkill
- legacy-to-agentic-estimateSkill
- mux-dispatchSkill
- non-functional-requirementsSkill
- persistent-memorySkill
- plan-reviewSkill
- playwright-test-implementationSkill
- policy-validationSkill
- pre-planning-analysisSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 6 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
6
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 16
When an existing ADR is incomplete (missing alternatives, consequences, or status)
- Distinctive: line 35
If alternatives are not yet documented:
- Distinctive: line 38
Standalone / interactive use: if the conversation already contains discussed alternatives, record those.
- Distinctive: line 102
After generating the ADR, confirm:
- Distinctive: line 115
All required sections are present: Context, Decision, Status, Alternatives Considered, Consequences
- Distinctive: line 117
No consequence is vague (no "will improve X" without a metric or qualifier)
“Common” is measured: kept by at least 6 in 10 of the 5 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
ADR Generator
Produces a complete Architecture Decision Record (ADR) for a significant design decision, with these sections: Context, Decision, Status, Alternatives Considered, Consequences. Uses the formal ADR format defined in the decision-writing skill. ADRs are generated inline during design doc creation and embedded under the ## Architecture Decision Records section.
When to Use
- During
k-design-doc-creation.sop.mdPhase 2 (creation) — triggered automatically for every significant design decision; Phase 3 uses this skill primarily to verify placement, not for full regeneration, but may still apply Step 1's automated fallback to fill in a still-missing Alternatives row - Standalone: engineer asks to document a specific decision as an ADR
- When an existing ADR is incomplete (missing alternatives, consequences, or status)
A decision is significant if it:
- Chooses between two or more viable architectural options
- Has non-trivial cost, latency, complexity, scalability, or operability implications
- Will be difficult or expensive to reverse
Steps
Step 1: Gather Decision Context
Collect from the current conversation or document:
- Decision context — what problem triggered this decision, what constraints apply
- Chosen option — the option selected and the primary reason
- Alternatives — at least 2 alternatives with pros and cons (use
trade-off-evaluatorscores if available) - Status — one of:
Proposed,Accepted,Deprecated,Superseded
If alternatives are not yet documented:
- Automated SOP workflow (e.g., invoked from
k-design-doc-creationPhase 3 — no user interaction allowed): infer at least one viable alternative from the decision context and available trade-off data (fromtrade-off-evaluator). Do NOT ask the engineer. If no credible alternative can be inferred, still produce an Alternatives table with the chosen option PLUS a placeholder row| <TBD — flagged for adversarial review> | — | — | — |, and flag the ADR for the adversarial-review phase. This ensures the Alternatives table always has ≥ 2 rows so the Quality Gate is satisfied. - Standalone / interactive use: if the conversation already contains discussed alternatives, record those. If no alternatives have been explored yet, suggest 1-2 plausible alternatives and discuss them with the engineer before generating the ADR — do not silently invent alternatives the engineer never considered.
Step 2: Generate ADR
Apply the formal ADR format from decision-writing:
### ADR-N: <Title>
**Status:** Accepted
#### Context
<What situation or problem triggered this decision. What constraints apply.
One paragraph. Specific — include data where available.>
#### Decision
<The decision in one sentence. Active voice. "We will use X because Y." — state the concrete technical reason for acceptance; crisp, specific, fact-checked.>
#### Alternatives Considered
| Option | Pros | Cons | Why Not Chosen (crisp, specific, fact-checked — 1–3 sentences) |
| ----------------- | ---- | ---- | ---------------------------------------------------------------- |
| Option A (chosen) | ... | ... | — (chosen; acceptance rationale in Decision above) |
| Option B | ... | ... | <concrete technical reason: limit / latency / cost / ops burden> |
| Option C | ... | ... | <concrete technical reason: limit / latency / cost / ops burden> |
_Scored trade-off matrix: see `<name>-tradeoffs.md` or inline table above._
#### Consequences
**Good:**
- <specific, measurable positive outcome>
**Bad:**
- <specific cost, risk, or constraint introduced>
**Neutral:**
- <process or tooling change with no net positive/negative>Step 3: Assign ADR Number
ADRs are numbered sequentially within the document. Check the existing ## Architecture Decision Records section for the highest existing ADR number and increment by 1. If no ADRs exist yet, start at ADR-1. Numbers are scoped to a single document — the same ADR-N in different design docs are unrelated decisions. When cross-referencing an ADR from outside its own document, name the source document (e.g., "see ADR-3 in payments-design.md") rather than the bare number.
Step 4: Embed in Document
Insert the ADR under the ## Architecture Decision Records section of the design doc. If the section does not exist, create it after the Implementation Plan section.
Cross-reference the ADR from the relevant section of the design doc (e.g., in the Solution Overview where the decision is first mentioned, add: _(See ADR-N)_).
Pitfalls
- Do not generate an ADR for implementation details — ADRs are for architectural choices, not for "which variable name to use" or "which library version to pin".
- Do not omit alternatives — an ADR with no alternatives is a record of a fait accompli, not a decision. Minimum 2 alternatives including the chosen option.
- Do not use vague consequences — "will improve performance" is not a consequence. "Expected to reduce p99 latency from 450ms to <200ms based on load test" is.
- Status must be set — default to
Acceptedfor decisions made during creation. UseProposedif the engineer has not yet confirmed the choice.
Verification
After generating the ADR, confirm:
- ADR number is unique within the document
- Context section is specific (includes data or constraints, not just "we needed to decide")
- Decision is one sentence in active voice
- At least 2 alternatives are listed with specific pros/cons
- Consequences are split into Good / Bad / Neutral with at least one entry each
- ADR is cross-referenced from the relevant section of the design doc
Quality Gate
The ADR passes when:
- All required sections are present: Context, Decision, Status, Alternatives Considered, Consequences
- Alternatives table has ≥ 2 rows (including the chosen option)
- No consequence is vague (no "will improve X" without a metric or qualifier)
- ADR is embedded in the document under
## Architecture Decision Records - Rationale quality bar met — every
Decision(acceptance) and everyWhy Not Chosencell (rejection) is crisp (1–3 sentences), technically specific (cites limits/latency/cost/ops), and fact-checked (AWS claims validated viaaws-service-validator). Seedecision-writingADR format for the full bar and example. Rows flagged<TBD — flagged for adversarial review>are exempt and deferred to the adversarial-review phase.
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.

