62d62ef4fd
The fixture chat panel is now a working agent. Ask Reginald a question and it consults the real project through a tool loop, then answers in grounded prose. Read-only v0 — writes still go through the propose-approve controls. core (@commitea/core/agent): - chat-client: OpenAI-wire chat completions over an injected fetch (same seam as gitea). Points at any OpenAI-compatible endpoint (LM Studio/Ollama/OpenAI). - model-router: small model for prose + the read tool; big model reserved for later decomposition (pickModel). - agent-loop: runAgentTurn drives call→tool→result→call until prose (or a step budget), recording each tool step. Injected complete + execute → fully testable. - query-project: the single read tool's engine — compact focus/board/calibration/ issue/search views built from scheduler + lifecycle + calibration; unbuilt views return a notImplemented marker (never fabricated). The model reports, never computes. - agent-tools: query_project declaration + Reginald's system prompt. app: - main model bridge (model:status, model:chat) runs the loop; query_project reconciles the repo and builds the view. Model traffic stays in main (token/CSP). gitea.ts refactored to share getGiteaClient + reconcileSnapshot. - preload + global.d.ts expose the model bridge; useChat drives the panel — real agent turn when a model is configured, scripted fixture reply otherwise (so fixture e2e is unchanged). A subtle "consulted the project" activity line. Model config (env, defaults to LM Studio on :1234): COMMITEA_MODEL_URL / _SMALL (google/gemma-4-e4b) / _BIG (qwen/qwen3.6-35b-a3b). COMMITEA_E2E=1 keeps it unconfigured so the panel stays scripted. Verified: 88 core tests green (14 agent: client parse, loop tool/error/budget, all views) + a gated live integration test. Desktop typecheck clean, 14 fixture e2e green. Gated live e2e drives the real app against gitea + gemma-4-e4b: asked "what now?", Reginald called query_project and answered "focus is on issue #2" (the real scheduler pick). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
CommiTea
AI project manager built on Gitea. A deterministic scheduler does the math
(Monte Carlo forecasts over your own estimate-vs-actual history); the agent —
Reginald — captures work via interview, negotiates priorities, and explains the
consequences. Gitea holds human-authored intent; a pm-state repo holds
machine-derived state. The LLM never does math.
Structure
apps/desktop— Electron app (electron-vite, React, Tailwind)packages/core— pure TypeScript: label schema, scheduler, gitea client (no Electron imports)docs/PLAN.md— product plan (goals, phases, not-doing list)docs/design/— design handoff: tokens, component contracts, interactive prototype (docs/design/ui_kits/app/index.htmlvia a static server), brand voice rules indesign_system_readme.md
Development
yarn # install
yarn dev # electron app, logs tee to apps/desktop/desktop.log
yarn test # unit tests (vitest)
yarn typecheck
End-to-end (Electron + Playwright)
yarn workspace @commitea/desktop e2e # build, then drive the built app
yarn workspace @commitea/desktop e2e:only # reuse existing out/ build (tight loop)
yarn workspace @commitea/desktop e2e:report # open the last HTML report
Tests launch the built app (out/main/index.js) through Playwright's
_electron API — no browser project, no chromium download. Fixtures and page
objects live in apps/desktop/e2e/; screenshots land in
e2e/.artifacts/screens/ for visual review. Page objects use user-facing
locators (getByRole/getByText), never CSS/DOM structure.
Conventions
- Yarn 4 workspaces; ESM everywhere;
.jsextensions on relative imports @commitea/corestays pure — unit-testable without Electron or network- Design tokens are the source of truth (
apps/desktop/src/renderer/src/design/, mirrored fromdocs/design/); Tailwind maps onto the CSS custom properties, never redefines them - Reginald's voice rules live in
docs/design/design_system_readme.md— no emoji, no point-date forecasts, wit in sentences never in buttons