Context

This is where Spec Kitty's shared vocabulary and governance model live: the canonical glossary contexts, the audience/persona catalog used across the docs, and the Charter governance model that shapes what your agent is told to do and why.

  • Glossary contexts (*.md) — canonical terminology per bounded context, relocated from docs/context/. These remain the doctrine-extraction source consumed by scripts/generate_contextive_glossaries.py; the dashboard glossary seed files under .kittify/glossaries/ are unchanged.
  • audience/ — architecture audience personas (internal/external), relocated from docs/context/audience/.
  • Charter-era overview — the current Spec Kitty 3.2 Charter governance model, distilled here for quick reference. See How Charter Works and the Governance Files Reference.

Glossary reference notes

Companion notes to the glossary contexts, folded in here from the legacy root-level glossary/ stub and the previously homeless context docs:


Spec Kitty 3.2 Charter-era overview

You are looking at the current Spec Kitty 3.2 documentation surface for new projects and upgrades.

Answer summary

  • Current target version: Spec Kitty 3.2.
  • Current runtime model: Charter-era missions with governed context injection.
  • Current governance source: .kittify/charter/charter.yaml (the git-tracked, authoritative file; charter.md is a human-readable companion the runtime never parses).
  • Current mission loop: spec-kitty next --agent <name> --mission <slug>.
  • Upgrade path: Migration to Spec Kitty 3.2.

What is Charter?

Charter is the governance layer introduced in Spec Kitty 3.x. .kittify/charter/charter.yaml is the single git-tracked, authoritative charter — it's what charter synthesize reads and what the runtime resolves at every governed mission action. .kittify/charter/charter.md is a human-readable narrative companion for authors; the runtime never parses it. The flow is:

charter.yaml  ->  charter synthesize  ->  validated Charter state  ->  governed prompt/context

When you run spec-kitty next --agent <name> --mission <slug>, the runtime automatically injects the relevant Charter context into the prompt file it returns for the next mission action. For standalone work, spec-kitty dispatch "<request>" uses the same governed-context model and records an Op trail.

For the full mental model, see How Charter Works.

Documentation by type

Tutorials — learning-oriented

How-to guides — task-oriented

Reference — authoritative specifications

Explanation — conceptual background

Migration

Upgrading from an earlier version? See Migrating from 2.x / Early 3.x — what changed, migration steps, and known failure modes.

What is archived

Documentation for Spec Kitty 1.x and 2.x is preserved through the migration hub for historical context. The 2.x governance model did not include the DRG-backed synthesis pipeline or the retrospective learning loop. If you are running a current project, use the 3.2 documentation above.