Skip to content

Core concepts

A quick tour of the vocabulary. This is a glossary, not a homework assignment: skim the nouns you bumped into, then follow the deeper links when one matters.

Agent harness

The layer around AI models that gives them work, context, tools, state, permissions, review rules, and recovery paths. A chat assistant answers a conversation. An agent harness runs a workflow.

Guildhall is an agent harness for local software projects.

Project / workspace

In the product, you mostly think in projects. In the runtime and config layers, the same unit is still often called a workspace.

Either way, it means one directory containing a ./guildhall.yaml and shared Guildhall project state under ./.guildhall/. Local transcripts and bulky run history live under ~/.guildhall/data/projects/. See Workspaces.

Task

A framed piece of work with a status, a domain, an optional blueprint/spec, acceptance criteria, hard gates, and review verdicts. Tasks move through a fixed lifecycle: proposed → exploring → spec_review → ready → in_progress → review → gate_check → done (terminal: done, shelved, blocked). See Task lifecycle.

Guildhall also sizes tasks. Tiny and small work can usually move as one unit; large work gets split recommendations; epic work becomes containing work with linked child tasks. The hierarchy is a link between tasks, not a special "parent" status.

Request shape

How Guildhall reads a New Request before it becomes work. The request might be a task, a question, a settings change, a repair, a broad intake, or an ambiguous "spec or implementation?" ask.

When the shape is ambiguous, Guildhall records the uncertainty and asks a focused question instead of guessing.

Component stack

The linked pieces inside a broader request: policy decision, documented spec, implementation surfaces, API or data changes, docs, release work, and verification. A component stack helps Guildhall keep related work connected without turning it into one oversized task.

Pressure-Test Intake

Guildhall's completeness pass before work becomes implementation. Every task is pressure-tested, but the depth scales with risk and uncertainty. Small work may only need automatic scope, verification, and review checks; broad or risky work can enter a one-question-at-a-time intake loop in Thread. See Pressure-Test Intake.

Blueprint

An accepted plan for a project, release, phase, or task. A blueprint names what is being built, why it matters, what is in and out of scope, how it will be verified, and what assumptions are still risky. See How Guildhall builds.

Framing

The work of turning broad intent into buildable structure: phases, task boundaries, dependencies, active tranches, attached questions, likely files, and verification standards. Framing is what prevents a project from becoming a pile of drafts.

Domain

A named slice of the project, such as ui, backend, or infra. Every task declares a domain. Work that crosses domains becomes an explicit handoff instead of a quiet surprise.

Driver

The source of product demand. A driver answers whose need decides what matters first. It can have path hints, but it is not just a folder. A customer-facing app might be the primary driver, a shared package might be the provider, and a Storybook or e2e surface might be the proof driver.

Delivery package

The thing Guildhall is trying to deliver. A delivery package can contain child tasks for implementation, docs, tests, proof, or cleanup. Hierarchy describes the shape of delivery; dependencies describe execution order.

Primitive

A foundational project piece downstream work uses or must respect. UI components and design tokens are primitives, but so are API clients, response envelopes, auth guards, data schemas, event buses, shared test harnesses, and workflow conventions.

Guildhall tracks which tasks use primitives and which tasks prove them. A primitive is not ready just because code exists; it needs proof or an explicit owner decision. See How Guildhall chooses work.

Structure

The advanced project map for repo domains, multi-project relationships, handoffs, and shared contract questions. Most local delivery work should be understandable through drivers, delivery packages, tasks, dependencies, primitives, and proof. If an advanced structure item needs your judgment, the discussion routes back to Thread.

Project graph

The advanced ownership map across projects, domains, packages, external references, and delivery channels. It matters when one registered project really needs something from another project or owner boundary. It should not be the mental model for ordinary local task dependencies.

Contract surface

A shared surface individual specs need to fit. Examples include a component API, endpoint family, event stream, schema, state machine, MCP resource, design system, or domain capability. In the default project model, these often show up as primitive invariants and proof expectations rather than as a separate thing you need to manage.

Owner input

A linked decision session for your judgment. Thread owns the conversation; Needs You owns the alert that something is waiting; Overview, Work, Structure, and Settings can link to the same decision without creating duplicate question cards.

Coordinator

The general-contractor role for a domain. Defined in ./guildhall.yaml with a mandate, concerns, and lists of autonomous decisions and escalation triggers. The coordinator decides whether tasks in its domain advance and keeps the local plan coherent; see Coordinators & domains.

Agent

A stateful, tool-using conversation. Five built-in roles: spec, coordinator, worker, reviewer, gateChecker. Each role maps to a model and provider choice, with machine-wide defaults and optional project-level overrides. See Agents & models.

Guild

The metaphor behind the product name. A guild hall is a shared place where different skilled trades coordinate work under common standards.

In Guildhall, a guild is a specialist persona with principles, a review rubric, and deterministic checks for relevant tasks. Examples: Accessibility Specialist, Color Theorist, Frontend Engineer. Guilds attach to tasks via an applicability predicate; multiple guilds can inspect one task (fan-out).

Change order

An explicit revision to the accepted plan. Use a change order when new evidence changes scope, sequencing, assumptions, or task shape. Change orders stay visible because silent drift destroys trust.

Git Story

The closure state for project or task changes: clean, dirty, local commits, no upstream, pushed, PR open, merged, local-only, deferred, conflict, or unknown. It is the bit that tells you whether "done" also means the repo is in an intentional state. See Git Story Closure.

Punch list

The small remaining finish work after the main build is coherent: cleanup, docs alignment, missing tests, release notes, polish, and deferred follow-ups. The punch list separates shipping blockers from nice-to-have finish work.

Lever

A named decision point, such as merge_policy, reviewer_mode, or worktree_isolation. Levers keep behavioral choices in plain YAML with a rationale, instead of hiding them as hardcoded defaults. See Levers and Onboarding and levers.

Hard gate / soft gate

Hard gates are deterministic checks a task must pass before it can complete: lint, typecheck, tests, custom shell commands. Soft gates are rubric items scored by a reviewer. Together they form the completeness bar.

Business envelope

The project-level Goals + Guardrails document that defines what Guildhall is allowed to do and what it is not. business_envelope_strictness controls enforcement mode.

Skill

A bundled instruction set an agent can invoke. Skills are how you teach Guildhall reusable procedures without baking them into prompts.

Hook

A user-defined command, prompt, HTTP call, or agent invocation that fires at lifecycle events (session_start, pre_tool_use, etc.). Hooks let you plug in audit loggers, external approvals, and custom side effects.

MCP server

A Model Context Protocol server — stdio, HTTP, or WebSocket — whose tools become agent-callable.

Session

A persisted snapshot of an agent's conversation: messages, model, usage, tool metadata. Enables warm resume — you can interrupt an agent and it will pick up from the last snapshot.

Released under the FLL-1.2 License.