culture-rules

Rules engine for the AgentCulture mesh: rules → conditions → workflows → actions, carried out by actors (agents, humans, code, services, robots, …). It is a Python backend library, culture_rules, plus a Node.js + React Flow visual editor with three tabs: Rules | Workflows | Actors.

Status: scaffold. This repo was provisioned from culture-agent-template. The agent-first CLI, mesh identity, skill kit and CI/publish baseline are in place. The engine, HTTP API and editor are planned and not built yet. The build brief is #1, and the product model and UX are #2.

The model (planned)

A rule says when work happens, and reads as a small flow:

PR approved  →  base = main  →  Review PR  →  Comment
  trigger        condition       workflow      action

Human actors make runs long-running and asynchronous, so the engine will persist run state rather than hold it in memory.

The editor (planned)

The editor has exactly three primary tabs:

• Rules shows each rule as a compact graphical flow, and grows new rules progressively.
• Workflows is the React Flow canvas. Inputs, outputs and variables are mapped visually.
• Actors is the registry of actors and their capabilities.

Run history and debugging appear in context, never as extra tabs. The design goal is a visual composition tool, not an admin dashboard:

• large type and large targets;
• minimal chrome and prose;
• smooth transitions;
• keyboard- and reduced-motion-accessible.

Quickstart (what works today)

uv sync
uv run pytest -n auto                 # run the test suite
uv run culture-rules whoami           # identity from culture.yaml
uv run culture-rules learn            # self-teaching prompt (add --json)
uv run teken cli doctor . --strict    # the agent-first rubric gate CI runs

CLI

Every command supports --json. Results go to stdout, and errors and diagnostics go to stderr; the two are never mixed. Exit codes: 0 success, 1 user error, 2 environment error, 3+ reserved. Rule, workflow and actor verbs will follow the same contract, with writes dry-run by default and --apply to commit them.

Prompt files by harness

Four harnesses read four root files, with no shared base. Each file is read by exactly one harness, and there is intentionally no AGENTS.md:

There are two separate selections over this clone:

• The interactive harness: whichever binary you run. All four are live at once.
• The mesh resident: the backend that culture.yaml declares, here claude.

See docs/harness-selection.md and docs/automation-contract.md.

Skills

.claude/skills/ vendors 19 skills (cite-don't-import), including the devague workflow (scope → think → spec-to-plan → assign-to-workforce), which drives the build. See docs/skill-sources.md for provenance and the re-sync procedure.

Contributing

See CLAUDE.md for the conventions:

• every PR bumps the version;
• PRs go through the cicd lane;
• worktree layout and memory discipline.

License

Apache 2.0 — see LICENSE.