From santhoshstack
Generates a human-readable architecture document for any codebase. Use when onboarding contributors, presenting to stakeholders, explaining how a system works to non-technical collaborators, or after a major architectural change. Outputs docs/ARCHITECTURE.md with a Mermaid system diagram, plain English narrative, data flow walkthrough, and tech stack rationale. Trigger on: "map the architecture", "explain how this works", "architecture doc", "system map", "explain to non-technical", "onboard contributor", "document the system", "draw the architecture".
How this skill is triggered — by the user, by Claude, or both
Slash command
/santhoshstack:arch-mapThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Generates a layman-readable architecture document from any codebase.
Generates a layman-readable architecture document from any codebase.
This is a communication artifact, not a resume artifact. It is human-facing —
shareable with stakeholders, contributors, and non-technical collaborators. It lives
at docs/ARCHITECTURE.md and is never auto-loaded by context-manager.
| Mode | When |
|---|---|
| GENERATE | No docs/ARCHITECTURE.md exists, or user wants to rebuild from scratch |
| UPDATE | File exists, architecture has changed, user wants to refresh it |
For UPDATE: read the existing file first. Change only what actually changed — surgical edits.
Read in this order. Stop when you can answer all five questions below.
README.md — what does this project say it does?package.json, requirements.txt, go.mod, Cargo.toml, pyproject.toml) — what tech is in use?main.*, index.*, app.*, server.*, routes.*git log --oneline -10 — what has been worked on recently?docs/context/ — if context-manager is set up, skim domain files for architecture notes and the reference file ({domain}-ref.md) for what/how/why/whereFive questions to answer before writing:
If question 5 is unclear for any component, write "—" in the Why column. Do not invent rationale.
Create or update docs/ARCHITECTURE.md using this exact structure:
# [Project Name] — Architecture
> Last updated: {YYYY-MM-DD}
> For AI session context, see `docs/context/`. This file is for humans.
## What this is
{1–2 sentences. Plain English. No jargon. What problem does this solve and who uses it?}
## System map
{Mermaid diagram — see rules below}
## How it works — step by step
Walk through the single most important user action (the one that touches the most components):
1. User {does what}
2. {Component A} receives the request and {does what}
3. {Component B} {does what}, querying {Component C} for {what}
4. Result returns to the user as {what the user sees}
Keep it to 4–8 steps. Plain English labels only.
## Tech stack
| Layer | Technology | Why |
|---|---|---|
| {layer} | {tech} | {one line — why this, not something else} |
## Key constraints
Non-obvious rules that shape every decision in this codebase:
- **{constraint name}**: {what it is and why it matters}
## What is not built yet
{Only include if there are known significant gaps. Skip this section if nothing notable is missing.}
graph TD (top-down) for most projectsgraph LR (left-right) if the flow is a clear pipeline (ingest → process → store → serve)subgraph to group related components: subgraph Frontend, subgraph Backend, subgraph DataSearchUI[Search UI] not SearchUI[pgvector HNSW]A -->|user query| Bsubgraph ExternalExample for a web app with search:
graph TD
subgraph Frontend
UI[Web App]
end
subgraph Backend
API[API Server]
Search[Search Engine]
end
subgraph Data
DB[(Database)]
VectorDB[(Vector Store)]
end
subgraph External
LLM[AI Provider]
end
UI -->|search query| API
API --> Search
Search --> VectorDB
Search --> DB
API -->|generate embeddings| LLM
LLM --> VectorDB
If docs/context/pointers.md exists, check if it already has an architecture link.
If not, add one line immediately after the project title/header:
> Architecture map: [`docs/ARCHITECTURE.md`](../ARCHITECTURE.md) — human-readable system overview
Do not modify anything else in pointers.md.
Tell the user:
Do not summarise the full document. One short paragraph is enough.
npx claudepluginhub isanthoshgandhi/santhoshstack --plugin santhoshstackGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.