CLI Forge Hld

Optimization: This skill uses on-demand loading. Heavy content lives in references/ and is loaded only when needed.

Language rule: Skill instructions are written in English. When generating user-facing output, detect the project's primary language (from README, comments, docs, commit messages) and produce the document in that language. If the project is bilingual, ask the user which language to use before proceeding.

Diacritics rule: When the chosen output language uses diacritics (French, Spanish, Portuguese, etc.), render them correctly — including on capitals (É, À, Ç, Ô…). Never strip accents to ASCII: write "sécurité" not "securite", "Rôle" not "Role". ASCII-flattened French is a defect, not a style.

Writing style: Favor clarity — short sentences, one idea per sentence, plain words. A metaphor only when it clarifies. Keep one consistent register (impersonal/infinitive for operational docs). Separate narrative prose, meant to make the reader understand, from telegraphic content (tables, commands, checklists) which stays terse for action.

No fabricated output: Never invent command output, console captures, API responses, metrics, or logs. When real output is needed, emit a clearly-marked placeholder (🚧 "replace with a real capture"), never plausible-but-fake text presented as real.

Forge HLD — High-Level System Design

"Everything is a trade-off. Why is more important than how." — Neal Ford & Mark Richards

Core Principles

1. Mitosis — the document grows only the sections the project actually needs. A CLI tool doesn't need a deployment diagram. A single-service app doesn't need a context map. Kill empty sections, don't fill them with boilerplate
2. Numbers before architecture — always estimate capacity before designing. The numbers determine whether you need 1 server or 100
3. Every decision needs a "why" — no technology choice without justification via ADR
4. Tradeoffs are explicit — use ATAM-style reasoning: "We chose X over Y because [quality attribute] matters more than [other attribute] in this context"
5. Diagrams are mandatory — use Mermaid for all visuals (C4, sequence, deployment)
6. Non-goals are as important as goals — they prevent scope creep (Google design doc tradition)
7. KISS — if a section can be one table instead of three paragraphs, use the table. If a diagram replaces a page of text, use the diagram. Readers scan, they don't read
8. Gotchas — read ../gotchas.md before producing output to avoid known mistakes

Workflow

Step 1 — Identify scope

Step 2 — Gather requirements

Extract or ask for functional requirements (use cases, actors, critical journeys) and non-functional requirements using ISO 25010:2023 as checklist:

If the user already provided context, use that instead of asking redundant questions.

Step 3 — Size the document (Mitosis)

S/M/L/XL semantics are canonical in ../shared/tiering.md — the tables below specialize them to system complexity and the arc42 section inclusion matrix.

Detect the tier from signals:

Section inclusion matrix:

Target document size:

Rules:

• Never include an empty section. If a section doesn't apply, omit it entirely — don't write "N/A"
• Merge small sections when possible. For tier S, Security + Data + API can be one "Technical Decisions" section
• State the tier at the top of the document so readers know the level of detail to expect
• If in doubt about whether a section is needed, ask: "Would removing this section cause someone to make a wrong decision?" If no, cut it

Step 4 — Run back-of-envelope estimations (if tier M+ or >1K users)

Read references/estimation-cheatsheet.md for formulas, unit conversions, latency numbers, and scalability laws. Compute traffic, storage, bandwidth, cache, and server estimates. Present results in a clear table — these numbers drive every subsequent design decision.

Step 5 — Write the HLD

Read references/sections.md for full section templates with Mermaid diagrams, tables, and examples. Only include sections selected by the mitosis matrix in Step 3. For tier S, merge related sections freely.

HLD Document Structure

# [System Name] — High-Level Design

## 1. Executive Summary
## 2. Goals and Non-Goals
## 3. Stakeholders & Constraints
## 4. System Context (C4 Level 1)
## 5. Solution Strategy
## 6. Container Architecture (C4 Level 2)
## 7. Capacity Estimations
## 8. Domain Boundaries (Strategic DDD)
## 9. Data Architecture
## 10. API Design (High-Level)
## 11. Security Architecture & Trust Boundaries
## 12. Deployment Architecture
## 13. Observability & SLOs
## 14. Failure Modes and Mitigation
## 15. Architecture Decision Records (ADRs)
## 16. Tradeoff Analysis
## 17. Risks and Open Questions
## 18. Glossary

HLD vs LLD Boundary

If you're writing column-level detail or class diagrams, you've crossed into LLD territory. Stop and suggest the user invoke /cli-forge-lld for that component.

Quality Scoring

After writing, score the HLD using the Architecture Completeness Index (ACI). Read references/scoring.md for the full formula, 12-item checklist with weights, thresholds, and derived metrics.

Anti-Patterns

Before delivering, verify the HLD avoids common traps. Read references/anti-patterns.md for the 10 anti-patterns table.

Review Checklist

All tiers: Tier stated | Goals + Non-Goals | C4 L1 | Solution strategy with rationale | 1+ ADR per decision | Arrows labeled | No empty sections M+: C4 L2 | NFRs addressed | Tech justified | Security model | Data flow (write + read path) L+: Capacity with formulas | Deployment architecture | Failure modes | SLIs/SLOs XL: Context map | ATAM tradeoff | Glossary Size: S: 2-4p | M: 5-10p | L: 10-18p | XL: 15-25p — no section exists just because it's in the template

Dynamic Handoffs

Rule: Recommend, don't auto-execute.

Integration with other cli-* skills

Reference Sources

• Kleppmann — Designing Data-Intensive Applications | Bass, Clements, Kazman — Software Architecture in Practice | Ford, Richards — Fundamentals of Software Architecture | Brown — The C4 Model | Evans — Domain-Driven Design
• arc42 | C4 Model | MADR | Google Design Docs | French DAT | ISO 25010:2023
• ByteByteGo | Software Engineering Radio | InfoQ | Software Engineering Daily