$npx -y skills add ncklrs/startup-os-skills --skill adrExpert Architecture Decision Record (ADR) creation and lifecycle management
| 1 | # Architecture Decision Records |
| 2 | |
| 3 | Expert guidance for creating and managing Architecture Decision Records based on Olaf Zimmermann's research and the MADR (Markdown Architectural Decision Records) methodology. |
| 4 | |
| 5 | ## How This Skill Works |
| 6 | |
| 7 | When invoked, this skill guides you through the full ADR lifecycle: |
| 8 | |
| 9 | 1. **Readiness assessment** — Evaluate whether a decision is ready to be recorded (START criteria) |
| 10 | 2. **Creation** — Write well-structured ADRs using the MADR template with anti-pattern avoidance |
| 11 | 3. **Quality evaluation** — Check completeness using the ECADR "Definition of Done" criteria |
| 12 | 4. **Lifecycle management** — Maintain ADR logs, supersede outdated records, track decision status |
| 13 | |
| 14 | ## Core Philosophy |
| 15 | |
| 16 | **"Don't decide too early, this harms flexibility. Don't decide too late either."** |
| 17 | |
| 18 | ADRs function as decision journals — not blueprints or policies. They capture the *rationale* behind architecturally significant choices: the problem context, the alternatives considered, the criteria applied, and the consequences accepted. Good ADRs are executive summaries that balance tradeoffs, not sales pitches for a predetermined outcome. |
| 19 | |
| 20 | ## When to Use This Skill |
| 21 | |
| 22 | - Creating a new ADR for an architectural decision |
| 23 | - Evaluating whether a decision is ready to be recorded |
| 24 | - Reviewing an existing ADR for completeness and quality |
| 25 | - Writing or customizing a MADR template for a project |
| 26 | - Assessing whether a requirement is architecturally significant |
| 27 | - Managing an ADR log (superseding, deprecating, linking decisions) |
| 28 | - Training teams on ADR best practices |
| 29 | |
| 30 | ## Command Modes |
| 31 | |
| 32 | | Argument | Action | |
| 33 | |----------|--------| |
| 34 | | `create` | Walk through creating a new ADR interactively | |
| 35 | | `review` | Evaluate an existing ADR against quality criteria | |
| 36 | | `check-ready` | Assess decision readiness using START criteria | |
| 37 | | `check-done` | Assess ADR completeness using ECADR criteria | |
| 38 | | `template` | Output a blank MADR template ready to fill in | |
| 39 | | `list` | Help organize and maintain an ADR log | |
| 40 | | *(no args)* | General ADR guidance based on context | |
| 41 | |
| 42 | ## The MADR Template (Quick Reference) |
| 43 | |
| 44 | ```markdown |
| 45 | # ADR-NNNN: [Short Title Describing Decision] |
| 46 | |
| 47 | ## Status |
| 48 | [proposed | accepted | deprecated | superseded by ADR-XXXX] |
| 49 | |
| 50 | ## Date |
| 51 | YYYY-MM-DD |
| 52 | |
| 53 | ## Decision Makers |
| 54 | [Who made or approved this decision] |
| 55 | |
| 56 | ## Context and Problem Statement |
| 57 | [1-3 paragraphs: What is the issue? Why does it matter? |
| 58 | Frame as a question when possible.] |
| 59 | |
| 60 | ## Decision Drivers |
| 61 | - [Driver 1: quality attribute, business constraint, or technical concern] |
| 62 | - [Driver 2] |
| 63 | - [Driver 3] |
| 64 | |
| 65 | ## Considered Options |
| 66 | 1. [Option A — the chosen option] |
| 67 | 2. [Option B] |
| 68 | 3. [Option C] |
| 69 | |
| 70 | ## Decision Outcome |
| 71 | Chosen option: "[Option A]", because [justification referencing decision drivers]. |
| 72 | |
| 73 | ### Consequences |
| 74 | - Good, because [positive consequence] |
| 75 | - Good, because [another benefit] |
| 76 | - Bad, because [accepted tradeoff] |
| 77 | - Bad, because [known limitation] |
| 78 | |
| 79 | ## Pros and Cons of Options |
| 80 | |
| 81 | ### Option A |
| 82 | - Good, because [advantage] |
| 83 | - Bad, because [disadvantage] |
| 84 | |
| 85 | ### Option B |
| 86 | - Good, because [advantage] |
| 87 | - Bad, because [disadvantage] |
| 88 | |
| 89 | ### Option C |
| 90 | - Good, because [advantage] |
| 91 | - Bad, because [disadvantage] |
| 92 | |
| 93 | ## Validation |
| 94 | [How will we verify this decision works? Code review, design review, |
| 95 | architectural fitness function, spike, etc.] |
| 96 | |
| 97 | ## More Information |
| 98 | [Links to related ADRs, RFCs, spikes, or external references. |
| 99 | Note confidence level and planned review date.] |
| 100 | ``` |
| 101 | |
| 102 | ## Key Frameworks |
| 103 | |
| 104 | ### START — Definition of Ready |
| 105 | Before writing an ADR, confirm all five criteria: |
| 106 | - **S**takeholders known and available |
| 107 | - **T**iming is the Most Responsible Moment |
| 108 | - **A**lternatives identified (minimum two) |
| 109 | - **R**equirements and context documented |
| 110 | - **T**emplate selected and instantiated |
| 111 | |
| 112 | ### ECADR — Definition of Done |
| 113 | Before marking an ADR as accepted, confirm all five criteria: |
| 114 | - **E**vidence that the design will work |
| 115 | - **C**riteria applied to compare at least two alternatives |
| 116 | - **A**greement from relevant stakeholders |
| 117 | - **D**ocumentation captured and shared |
| 118 | - **R**ealization and review plan scheduled |
| 119 | |
| 120 | ### ASR Test — Architectural Significance |
| 121 | A requirement warrants an ADR when it scores on these criteria: |
| 122 | - High business impact or risk |
| 123 | - Critical stakeholder concern |
| 124 | - Quality-of-service deviation from current architecture |
| 125 | - External dependency that is unpredictable or uncontrollable |
| 126 | - Cross-cutting concern affecting multiple components |
| 127 | - First-of-a-kind implemen |