Glossary

This glossary defines key terms used across the repository. Each entry includes a short definition and a concrete path-oriented example.

Core concepts

• •

  module

  • •A unit of code and configuration under /modules/<module_id>/ with its own strategy, sources of truth, and implementations.
  • •Example: /modules/homework_system/MANIFEST.yaml
• •

  module_type

  • •A classification used to organize modules and define allowed relationships and expectations.
  • •Example: /type_graph.yaml (when present) defines module types.
• •

  module_instance

  • •A concrete instantiation of a module under /modules/<module_id>/.
  • •Example: /modules/<module_id>/AGENTS.md defines the module’s local strategy.
• •

  ability

  • •The authorized execution unit for side effects, executed via the Tool Runner task interface.
  • •Example: abilities appear in ABILITY.md catalogs and are registered under /.system/registry/.
• •

  low-level ability

  • •A small, focused operation (script/api/mcp) that performs a specific side effect.
  • •Example: /.system/registry/low-level/<operation_key>.yaml
• •

  high-level ability

  • •An orchestrated workflow/agent that composes lower-level operations into a safe, repeatable flow.
  • •Example: /.system/registry/high-level/<ability_id>.yaml
• •

  agent

  • •A high-level ability that implements a broader workflow as an agent-like behavior.
  • •Example: a registry entry with type: agent surfaced into an ABILITY.md catalog.
• •

  hook

  • •Event-driven logic that runs on lifecycle events to provide guardrails and automation.
  • •Example: hook strategy and (future) registrations live under /.system/hooks/.
• •

  Agent Runtime

  • •The long-lived process that manages sessions, events, and AI turn control.
  • •Example: lifecycle and boundaries are documented in /.system/skills/ssot/repo/architecture-core-mechanisms/runtime-model/SKILL.md.
• •

  Tool Runner

  • •The component that executes abilities via task operations (create/run) and triggers hook gates.
  • •Example: ability execution workflow is documented in /.system/skills/ssot/repo/architecture-core-mechanisms/ability-workflow/SKILL.md.
• •

  Hook Runner

  • •The component responsible for running hooks for lifecycle events.
  • •Example: hook lifecycle workflow is documented in /.system/skills/ssot/repo/architecture-core-mechanisms/hook-lifecycle/SKILL.md.
• •

  workdocs

  • •Scenario-local task state used to record plans, context, checkpoints, and outcomes.
  • •Example: /modules/integration/workdocs/ or workdocs/active/T-YYYYMMDD-slug/
• •

  SSOT docs

  • •Stable, reusable documentation stored inside skill packages under /.system/skills/ssot/**.
  • •Example: /.system/skills/ssot/repo/architecture-core-mechanisms/repo-architecture-overview/SKILL.md
• •

  registry

  • •A machine-readable catalog that allows tools to discover and index entities (abilities, hooks, derived module indexes).
  • •Example: /.system/registry/ (abilities) or /modules/overview/ (derived module registries)
• •

  skill

  • •A provider-discoverable guidance bundle (progressive disclosure) that links to deeper SSOT docs and patterns.
  • •Example: /.system/skills/ssot/repo/execution-plans/execution-plan-authoring/SKILL.md
• •

  provider skill wrapper

  • •A generated provider-specific entrypoint that redirects the model to a single SSOT skill package.
  • •Example: /.codex/skills/execution-plan-authoring/SKILL.md
• •

  ability routing

  • •Routing to abilities using ABILITY.md and registries.
  • •Example: /modules/integration/ABILITY.md surfaces abilities whose registries include entrypoints: ["/modules/integration"].
• •

  scope

  • •A stable taxonomy label used to organize knowledge/skills/workflows.
  • •Example: scope: architecture in /.system/skills/ssot/repo/architecture-core-mechanisms/repo-architecture-overview/SKILL.md front matter.

Related

• •/.system/skills/ssot/repo/architecture-core-mechanisms/repo-architecture-overview/SKILL.md
• •/.system/skills/ssot/repo/architecture-core-mechanisms/knowledge-metadata/SKILL.md