architect-boot

SkillDocs & knowledge

Loads your project's past architecture decisions and helps record new ones using the Architect Claude Skill.

Use architect-boot in Claude, ChatGPT or Ahel Desktop

Free. Sign in, add architect-boot and connect your AI. About a minute.

Also: Claude Code · Cursor · Codex

Then ask your AI: use the architect-boot skill

Details

Instructions available. Your AI can read the instructions. Execution depends on the setup they require.

Add Ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

architect-bootStart free
About this skill

Use when architecture work starts or a tech-stack or pattern decision emerges, injects the project ADR index (.adlc/memory/adr/) as session context and pairs the decision with capture via /architect-specify; invoked from team-boot's Class Boots catalog.

What this skill tells your AI

The instructions your AI receives, as published by tikalk/adlc-team-skills in skills/architect/architect-boot/SKILL.md and read by Ahel’s review.

Overview

One of the five class boots surfaced by team-boot's Class Boots catalog. team-boot injects the always-relevant team context (constitution, CDR index, skills registry) at session start; this skill loads the architecture decision layer on demand — the Accepted ADR index from project memory — and pairs it with decision capture so new architecture choices are recorded (/architect-specify) rather than evaporating.

Reading decisions and recording decisions are one loop: existing ADRs inform the work; decisions made during the work flow back into the same record system.

When to Use

Invoke when:

  • Starting architecture work (system design, AD.md generation, cross-subsystem refactoring, dependency changes).
  • The session turns architectural mid-flight: a tech-stack choice, pattern selection, or "we chose X over Y" moment emerges.
  • Reviewing or validating an architecture and you need the decision backdrop.

For technology selection, also invoke tech-radar-boot — radar context should inform the ADR.

Core Process

Step 1: Locate the ADR Index

From the current working directory (do NOT walk up parent directories):

  1. Primary: docs/adlc/memory/adr/adr.md (generated /architect-implement; rows start | ADR)
  2. Fallback: .adlc/memory/adr/adr.md (legacy layout, pre-ADR-401)

If neither exists but docs/adlc/memory/adr/ADR-*.md or legacy .adlc/memory/adr/ADR-*.md files do, synthesize a lean table from each file (ID from filename; Sub-System/Decision/Status from frontmatter or first heading). If the directory is empty or absent, report 0 ADRs — never fabricate rows.

Step 1b: Read the ADR Drafts Index

Check .adlc/drafts/adr/ for ADR-*.md files with status: proposed in frontmatter. These are draft ADRs pending clarification. Collect ID / Title / Type / Status / Date from each file. If the directory is empty or absent, report 0 pending drafts.

Step 2: Inject ADR Context (Output Contract)

Emit before the task answer:

## ADR Context

| ID | Sub-System | Decision | Status |
|--|--|--|--|
| ADR-325 | lanes | Workflow lanes over flat queue | Accepted |

_Searched N ADRs, K matched._

## Drafts Pending Review

| ID | Title | Type | Status | Date |
|----|-------|------|--------|------|
| (from .adlc/drafts/adr/) |

_N pending drafts — run /architect-clarify to review._
  • Render ID / Sub-System / Decision / Status from the index (the full index also carries Date, Decision Makers, File — read the individual ADR-*.md when a task matches a row).
  • N = total index rows; K = rows relevant to the current task. K MUST equal the table rows shown — 0 matched → empty table, never inflate.

Step 3: Capture Architecture Decisions

TriggerAction
Tech stack choice, "we chose X over Y"ADR → direct write to .adlc/drafts/adr/ (pull tech-radar-boot context first)
Pattern selection, structural refactorADR → direct write to .adlc/drafts/adr/
ADR-class decision already in the ledgerverify capture happened; if not, re-surface

Add/refresh rows in the Session Decision Ledger (ID | Name | Type | Rel | Status | Clarify) for every ADR-class decision detected this session — including ones from before this boot was invoked. Mirror each decision as a task-list todo (draft → /architect-clarify at session end); after code-modifying tasks, add a trailing todo to sweep the ledger until Unrecorded: 0 pending · Unclarified: 0 drafts (a draft leaves Unclarified only via its clarify skill or an explicit user handoff to a named clarify or execute skill). At session end, deliver the clarify prompt naming each captured ADR draft in .adlc/drafts/adr/ (ID + skill); if the user defers clarify, mark those rows handed off.

Failure Handling

  • Missing index + missing records → emit the empty table with _Searched 0 ADRs, 0 matched._ and continue the user's task; never block.
  • Unparseable index rows → skip malformed rows, note the skip count.

Red Flags

  • Fabricating ADR rows or inflating K beyond the table shown.
  • Injecting the index but ignoring capture — the pairing is the point.
  • Walking up parent directories to find .adlc/.
  • Tech-selection ADRs recorded without radar context when the radar is available.

Verification

  • ADR Context table emitted with _Searched N ADRs, K matched._ (K = table rows).
  • Session Decision Ledger updated with detected ADR-class decisions.
  • Tech-selection decisions routed through tech-radar-boot → /architect-specify.

Signals

GitHub stars
141
Forks
2
Last commit
Oct 2026
Advanced
Item type
skill
Key
architect-boot
Source
github.com/tikalk/adlc-team-skills