engineering-docs) is an open-source npm plugin that transforms any AI coding agent into a principal-level documentation engineer. Instead of asking your agent to write a single document and hoping it knows the right questions to ask, Engineering Docs installs 22 composable, auto-triggering skills that guide your agent through the full software lifecycle — from the first rough idea all the way to deployment runbooks and incident postmortems. The plugin works across 14+ agent platforms and produces cross-consistent, industry-standard documents that a developer — or another AI agent — can pick up cold and build from.
Why Engineering Docs?
Coding agents are powerful, but they start every conversation without knowing your project’s architecture, your users, or your constraints. Engineering Docs solves this by equipping your agent with the same structured thinking a principal engineer applies before a line of code is written.Turn ideas into blueprints
Go from a business concept to a complete technical specification, architecture, and deployment plan — automatically.
Ask the right questions
Tool-call interviews surface exactly what each document needs. No repeated questions, no conversation pollution.
Industry-standard output
ISO/IEC/IEEE 29148, C4 Model, STRIDE, OpenAPI 3.1, Google SRE — every document follows the spec your team expects.
Works everywhere
Claude Code, Copilot, Cursor, Gemini CLI, Goose, Kimi Code, Cline, and 7+ more platforms. One install command each.
The 22 Skills
Every skill is an independent, callable unit with its own interview process, template, and quality gate. The orchestrator (using-engineering-docs) routes between all 22 skills automatically — you never need to know a skill’s name to get its output.
Discovery & Planning
using-engineering-docs
Orchestrator. Takes a raw idea and routes to all other skills automatically. Start here for any new project.
business-concept
Problem statement, target users, value proposition, monetization model, and project constraints.
project-plan
Scope, milestones, RACI matrix, timeline, and a full work breakdown structure.
user-personas-behavior
User personas, jobs-to-be-done, success metrics, and analytics instrumentation plan.
Specification & Feasibility
technical-specification
Functional and non-functional requirements in SRS/TSD format with EARS syntax and a full traceability matrix.
technical-feasibility-study
Go/no-go recommendation with evidence for any technically uncertain part of your idea.
Architecture & Design
system-architecture-document
C4 diagrams, 4+1 views, tech stack selection, and non-functional requirements.
architecture-decision-record
Immutable ADR log in MADR format — one record per significant technical decision.
database-design-document
ERD, normalized schema, indexing strategy, and a versioned migration plan.
api-design-document
REST/OpenAPI 3.1 contract, resource definitions, auth scheme, versioning, and RFC 7807 error shapes.
admin-access-control-specification
RBAC permission matrix, audit logging requirements, and break-glass access procedures.
technical-blueprint
Google/Stripe-quality technical design document per feature — one blueprint per non-trivial component.
ux-flow-specification
User journeys, screen-by-screen flows, and every UI state including edge cases and error states.
design-system-specification
Design tokens, component library, spacing system, and accessibility compliance requirements.
Quality & Risk
security-threat-model
STRIDE-based threat analysis, attack surface map, risk register, and ordered mitigations.
test-strategy-document
Testing pyramid, mocking strategy, CI quality gates, and coverage targets by layer.
implementation-plan
Dependency-ordered build sequence with phase gates, so developers start in the right order.
Delivery & Operations
deployment-plan
Release strategy, environment promotion flow, go/no-go gate criteria, and rollback procedures.
slo-error-budget-document
SLI/SLO targets, error budget policy, and burn-rate alert thresholds following Google SRE patterns.
technical-runbook
On-call operations manual with step-by-step procedures for every known failure mode.
disaster-recovery-plan
RTO/RPO targets, backup strategy, failover runsheets, and annual DR test schedule.
incident-postmortem
Blameless root-cause analysis using Five Whys — generated reactively after any production incident.
Custom Agents
Engineering Docs also ships four custom agent definitions — pre-configured personas your agent platform can invoke directly for focused tasks without running the full orchestration pipeline.Multi-Agent Compatibility
Engineering Docs installs natively on 14 agent platforms. The same 22 skills work identically regardless of which agent runs them.Agent config files (
AGENTS.md, CLAUDE.md, GEMINI.md, COPILOT.md, GOOSE.md, PI.md) use safe-write — they are only created if they don’t already exist at the destination, so your existing customizations are never overwritten.Philosophy
Engineering Docs is built on five principles that shape every skill’s design:- Systemic over ad-hoc — Rigorous, reproducible processes yield safer, cleaner software than one-off documents.
- Traceability — Every requirement links to a business goal and a test case. Nothing floats free.
- Visual-first — Complex architectures are mapped with Git-trackable Mermaid diagrams, not prose descriptions alone.
- Operational safety — No feature is complete without a deployment runsheet, monitoring plan, and rollback procedure.
- Blameless learning — Production failures are data points for system hardening, not blame assignments.
Next Steps
Quickstart
Install in under two minutes and run your first skill.
How It Works
Understand the orchestrator, interview mechanism, and six-phase pipeline in detail.