Environment for NeSy-Agentic Integrated Development
PanLL (pronounced "parallel") is a Human-Things Interface (HTI) designed to facilitate a "Binary Star" co-orbit between a Human Operator and a Neurosymbolic Machine.
This is not a passive text editor. It is an active, synchronous environment that manages Cognitive Ergonomics and kills Inertia.
Pane-L (Symbolic) |
Strict logic, types, and constraints. "The Law" that governs neural inference. |
Pane-N (Neural) |
Streaming "Inference Manifold" showing the Agent’s internal monologue and OODA loop. |
Pane-W (World) |
Central shared canvas (The Barycentre) where results manifest. |
-
Anti-Crash Library: Logical Circuit Breaker preventing unvalidated neural output from reaching the Barycentre
-
Vexometer: Real-time friction/cognitive load monitoring with anti-inflammatory UI adjustments
-
Contractiles: Adaptive state contracts between Operator and Machine
-
Feedback-O-Tron: Community-driven performance reporting and constraint suggestions
Frontend |
ReScript + Custom TEA (The Elm Architecture) |
Backend |
Rust + Tauri 2.0 |
Runtime |
Deno (tests, build orchestration) |
Testing |
Deno.test (97 JS tests) + cargo test (12 Rust tests) |
Styling |
Tailwind CSS 4.x |
Middleware |
Elixir/BEAM ( |
Verification |
Echidna (planned) |
PanLL uses Deno for runtime, testing, and build orchestration. ReScript compilation uses the symlinked compiler in node_modules.
# Compile ReScript modules
npx rescript build
# Watch ReScript sources during development
npx rescript build -w
# Run tests (97 JS tests + 12 Rust tests)
deno task test
# Watch mode for tests
deno task test:watch
# Development mode (Tailwind + static dev server + Tauri hot reload)
deno task dev
# Build Tailwind CSS
deno task css:build
# Full production build
deno task build-
TEA Architecture Guide — Comprehensive documentation for The Elm Architecture (TEA) implementation including Tea_Cmd, Tea_Sub, Tea_Vdom, and Tea_App modules with testing guide and best practices
PanLL can ingest the PanLL event-chain JSON exported by panic-attack and
surface it in Pane-W for MVP timeline review. Paste the JSON export into the
Event Chain panel and click Import JSON, or use Load File to open a
saved export.
The optional BEAM service (beam/panll_beam) supports all three API styles as
runtime-selectable options:
-
HTTP via Bandit/Plug (
/healthz,/v1/status) -
GraphQL via Absinthe (
/graphql,/graphiql) -
gRPC via
panll.v1.StatusService/GetStatus
Select protocols with PANLL_BEAM_APIS:
# defaults to all in non-test env
export PANLL_BEAM_APIS="http,graphql,grpc"For containerized runtime orchestration (Chainguard base + Cerro Torre pack/verify + selur-compose stack), see runtime/README.adoc.
deno task dev starts a static server for public/ on http://localhost:8000
to satisfy the Tauri devUrl.
|
Toggle Pane-L (Symbolic Mass) |
|
Toggle Pane-N (Neural Stream) |
|
Toggle Pane-W (Task Barycentre) |
|
Toggle Pane-W (alternate binding) |
v0.1.0-alpha (92% complete)
✅ Complete TEA implementation (Model-Update-View) ✅ All UI components functional (PaneL, PaneN, PaneW, Vexometer, FeedbackOTron) ✅ Tauri 2.0 backend with 8 working commands ✅ 109 tests passing (97 JS via Deno.test + 12 Rust via cargo test) ✅ npm→Deno migration complete ✅ Full ReScript compilation (180ms, 0 warnings) ✅ Anti-Crash token gating + backend validation hooks ✅ Event-chain import from panic-attack (paste + file) with persistence ✅ Security tool integration (panic-attack ambush, timeline import) ✅ Feedback report types wired to backend ✅ State persistence (localStorage auto-save) ✅ ARIA accessibility attributes across all components ✅ Keyboard shortcuts wired (Ctrl+Shift+L/N/B/W + vexation polling) ✅ Tailwind CSS 4.x build pipeline (Deno) ✅ BEAM API scaffold (HTTP/GraphQL/gRPC)
Next: v0.1.0 release preparation