First run

You've run npx @engineering11/e11-foundry install. Now what?

Everything is a skill

The installer wrote one directory per skill into your AI tool's skill directory (.claude/skills/, .agents/skills/, or .opencode/skills/). There is no separate "agent format" — agents, lifecycle workflows, and anytime tools all install the same way, and you invoke any of them directly by name.

Support files landed under .foundry/:

.foundry/
├── config.toml              # central config, team scope (installer-owned)
├── config.user.toml         # central config, user scope (gitignored)
├── custom/                  # your overrides — never touched by the installer
├── scripts/                 # resolve-customization.mjs, resolve-config.mjs, …
├── _config/
│   └── e11-help.csv         # assembled catalog of every installed skill
└── foundry/                 # installed module content + runtime config.yaml

You never edit installed skill files or the installer-owned config. Anything you want to change — personas, menus, templates, paths, review layers — is data you override from .foundry/custom/. See Customization.

The eight agents

Eight of the installed skills are agents — personas with menus. Invoke the skill, the skill becomes the agent.

Agent	Skill name	What it's for
E11 Master	e11-master	The entry point. Routes to everything else
E11 Bootstrapper	e11-bootstrapper	New repos, microservices, and SDKs
Analyst	e11-agent-analyst	Research, briefs, PRFAQ
PM	e11-agent-pm	PRD, epics and stories, readiness, course correction
UX Designer	e11-agent-ux-designer	UX + UI strategy
Architect	e11-agent-architect	Architecture spine, SDK composition
Dev	e11-agent-dev	Stories, quick flow, unattended runs, code review
Tech Writer	e11-agent-tech-writer	Docs as the deliverable

Full reference: Agents quick-reference.

Pick your ceremony — the loop is the default

The default way to ship with Foundry is the tight loop: intent → e11-dev-flow (or e11-spec-flow first when you want the contract reviewed before code) → layered self-review → commit. Everything heavier is a deliberate step UP the dial for higher-stakes work, not the standard path. From least to most:

You have…	Run	What happens
A tiny, zero-blast-radius change	e11-dev-flow	Its routing step picks oneshot mode — implement, review, commit in one condensed pass; no spec file, no board sync
A small, well-scoped change	e11-dev-flow	The full quick engine: clarify → spec → implement → layered self-review
A feature that needs a real requirements doc	e11-prd	The PRD fast path: batched questions, [ASSUMPTION]-tagged draft you react to, stakes-calibrated depth
A product that needs the whole treatment	Phases 1–4	Brief → PRD → UX → architecture → epics → story cycle. See the phase map
Work you'd rather not supervise	e11-dev-auto	One unattended iteration; HALTs done or blocked, never asks. See Unattended runs

The loop end of the dial is covered in depth in Quick flow. Starting a brand-new repo instead? That's e11-bootstrapper — see Start a new backend repo.

Start with E11 Master

When you don't know which skill you need, invoke e11-master. It greets you, loads .foundry/_config/e11-help.csv — the live catalog of everything installed — and presents it grouped by phase. Reply with a number, a code, or a description; Master dispatches the matching skill or hands the session to the right agent.

You: e11-master

Master: 🧭 Hi — E11 Master here. e11-help ([BH]) is always available.

        Anytime
          1. [BH]  e11-help       — Get unstuck; suggests next steps
          2. [DF]  Dev Flow       — Quick build/fix/refactor
          ...
        Phase 2 — Planning
         14. [PR]  PRD            — Create / Update / Validate (required)
        ...

You: I want to add a /healthz endpoint to one of our backend services.

Master: Small, established pattern — dispatching skill:e11-dev-flow.

You can also skip Master entirely: every skill is directly invokable (skill:e11-prd, skill:e11-code-review, skill:e11-knowledge sdk:auth, …). Agents exist for when you want a persona and a menu; skills exist for when you already know what you want.

When you're stuck: e11-help

e11-help ([BH]) is available from every agent and directly. Give it a sentence about what you're trying to do — it reads the installed catalog and your existing artifacts, figures out where you are in the lifecycle, and recommends the next skill (offering to run it for you when one is the clear next step).

You: skill:e11-help "I want to add a webhook endpoint to an existing service — where should I start?"

Where things land

• Planning artifacts (briefs, PRDs, UX, architecture, epics) → .foundry-output/planning-artifacts/
• Implementation artifacts (sprint status, stories, specs, reviews, retros) → .foundry-output/implementation-artifacts/
• Long-term project knowledge (research, document-project output) → docs/
• Your customizations → .foundry/custom/ (commit the team layer; .user.toml files stay personal)

All three artifact roots are the paths you chose at install time; .foundry-output and docs are the defaults.

Command codes you'll see often

Every catalog entry has a short letter code shown in menus. The codes are stable across installs.

Code	Skill / command	Where
BH	e11-help — what should I do next?	everywhere
QK	e11-knowledge — query the alloy corpus	every agent menu
CB / PF	Product Brief / PRFAQ	Analyst
MR / DR / TR	Market / Domain / Technical Research	Analyst
PR	PRD (Create / Update / Validate — one code, three intents)	PM
UX	UX Design	UX Designer
CA / IR	Architecture / Implementation Readiness	Architect
CE	Create Epics and Stories	PM
SP / SS	Sprint Planning / Sprint Status	direct
CS / VS	Create / Validate Story (e11-create-story)	direct
DS / CR	Dev Story / Code Review	Dev
SF / DF / DA	Spec Flow / Dev Flow / unattended Dev Auto	direct or Dev
QA / CP	E2E test generation / Checkpoint Preview	direct
ER / CC	Retrospective / Correct Course	direct
NB / NF / NS / RI	New Backend / Frontend / SDK / Resume Init	Bootstrapper
NM / AM	New / Augment Microservice	Bootstrapper
AR / ECH / VG / AA	The four review finders	anywhere
RF / MP / FI / DI	Refine / Multi-perspective / Forge Idea / Discovery	anywhere
CF	e11-customize — author overrides conversationally	anywhere

Full installed catalog: .foundry/_config/e11-help.csv in your project (source: src/foundry/module-help.csv).

Next

→ Concepts — the primitives: skills, agents, customization, alloy.