quorus-ai 0.5.0__tar.gz

quorus_ai-0.5.0/.dockerignore

@@ -0,0 +1,14 @@
+__pycache__
+*.pyc
+.pytest_cache
+.ruff_cache
+.git
+.gitignore
+tests/
+CLAUDE.md
+docs/internal/
+CONTRIBUTING.md
+.claude/
+.mcp.json
+messages.json
+uv.lock

quorus_ai-0.5.0/.env.example

@@ -0,0 +1,29 @@
+# Murmur Relay — Environment Variables
+# Copy to .env and fill in values, then: docker-compose up
+
+# Required: Shared secret for legacy auth (agents without API keys)
+RELAY_SECRET=change-me-to-a-long-random-string
+
+# Required: Secret for signing JWTs (min 32 chars)
+JWT_SECRET=change-me-to-a-long-random-secret-min-32-chars
+
+# Required: Bootstrap secret for creating the first tenant
+BOOTSTRAP_SECRET=change-me-bootstrap-secret
+
+# Optional: Server port (default 8080)
+PORT=8080
+
+# Optional: Postgres (defaults to local docker-compose postgres)
+# DATABASE_URL=postgresql+asyncpg://user:pass@host:5432/murmur
+
+# Optional: Redis (defaults to local docker-compose redis)
+# REDIS_URL=redis://host:6379/0
+
+# Optional: Logging level (DEBUG, INFO, WARNING, ERROR)
+LOG_LEVEL=INFO
+
+# Optional: Message limits
+MAX_MESSAGES=1000
+MESSAGE_TTL_SECONDS=86400
+RATE_LIMIT_MAX=60
+MAX_ROOM_MEMBERS=50

quorus_ai-0.5.0/.env.prod.example

@@ -0,0 +1,26 @@
+# Production environment template
+# Copy to .env.prod and fill in values
+
+# Required secrets (generate with: openssl rand -hex 32)
+JWT_SECRET=your-jwt-secret-min-32-chars-here
+BOOTSTRAP_SECRET=your-bootstrap-secret-here
+RELAY_SECRET=your-relay-secret-here
+POSTGRES_PASSWORD=your-strong-postgres-password
+REDIS_PASSWORD=your-strong-redis-password
+
+# Optional: Redis memory limit (default 512mb)
+# REDIS_MAXMEMORY=1gb
+
+# Optional: Relay settings
+# PORT=8080
+# LOG_LEVEL=WARNING
+# MESSAGE_TTL_SECONDS=86400
+# VISIBILITY_TIMEOUT_SECONDS=60
+# IDEMPOTENCY_TTL_SECONDS=300
+# RATE_LIMIT_WINDOW=60
+# RATE_LIMIT_MAX=60
+
+# Sentry — set in production via `flyctl secrets set SENTRY_DSN=...`.
+# Leave unset locally; init_sentry() is a no-op without a DSN.
+# SENTRY_DSN=
+# SENTRY_TRACES_SAMPLE_RATE=0.05

quorus_ai-0.5.0/.gitignore

@@ -0,0 +1,32 @@
+__pycache__/
+*.py[cod]
+*.pyo
+*.egg-info/
+dist/
+build/
+.eggs/
+.pytest_cache/
+.ruff_cache/
+.venv/
+.env
+messages.json
+.claude/
+uv.lock
+.DS_Store
+ROADMAP.md
+docs/launch/
+.mcp.json
+.murmur/
+.vercel
+*.mcp.*.json
+website/vite.config.d.ts
+website/vite.config.js
+website/*.tsbuildinfo
+MagicMock/
+
+# Graphify local knowledge graph (do not commit)
+# Share only the readable graph report and page; rebuild the rest with `graphify update .`
+graphify-out/*
+!graphify-out/GRAPH_REPORT.md
+!graphify-out/graph.html
+.graphify/

quorus_ai-0.5.0/.graphifyignore

@@ -0,0 +1,96 @@
+# --- generated by the graphify sweep, 2026-08-28 ---
+# Mirrors Rokan's .graphifyignore. The secrets block at the bottom matters
+# only once a semantic pass runs: the free AST pass reads symbol names,
+# but a semantic pass reads file CONTENT and sends it to a model.
+
+# Design assets (not code)
+design/
+*.html
+*.pdf
+*.png
+*.jpg
+*.jpeg
+*.svg
+*.gif
+*.mp4
+*.mov
+
+# Dependencies
+node_modules/
+__pycache__/
+.venv/
+venv/
+target/
+Pods/
+vendor/
+
+# Build outputs
+dist/
+build/
+.next/
+.expo/
+out/
+DerivedData/
+
+# Lock files
+package-lock.json
+yarn.lock
+pnpm-lock.yaml
+Cargo.lock
+Podfile.lock
+poetry.lock
+uv.lock
+
+# Generated
+*.tsbuildinfo
+*.generated.ts
+*.generated.js
+*.pb.go
+*_pb2.py
+
+# Test fixtures (boilerplate, low signal)
+**/__fixtures__/
+**/fixtures/
+**/snapshots/
+
+# Cache
+.cache/
+.parcel-cache/
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+
+# Git
+.git/
+
+# Scratch
+memory/
+scratch/
+experiments/
+tmp/
+
+# --- secrets and credentials: never into a graph, never to a model ---
+.env
+.env.*
+*.pem
+*.key
+*.p12
+*.keystore
+*.jks
+*.mobileprovision
+secrets/
+credentials/
+*credentials*.json
+service-account*.json
+GoogleService-Info.plist
+google-services.json
+*.sqlite
+*.db
+
+# --- patient / personal data, by convention ---
+# No real PHI is believed to live in these repos. Listed anyway, because a
+# convention only protects you if it predates the file that needs it.
+phi/
+patient_data/
+*patients*.csv
+*patients*.json

quorus_ai-0.5.0/.hatchignore

@@ -0,0 +1,13 @@
+.murmur/
+*.db
+*.db-shm
+*.db-wal
+dist/
+.venv/
+__pycache__/
+*.pyc
+.github/
+tests/
+website/
+docs/
+*.md

quorus_ai-0.5.0/AGENTS.md

@@ -0,0 +1,13 @@
+# AGENTS.md
+
+Read CLAUDE.md: it holds the project rules, test and lint commands for every agent.
+
+The essentials, for tools that don't read CLAUDE.md:
+
+- Python lives in `~/dev/Quorus/.venv`. Don't look for another one, and never
+  search the whole disk (`find /`) to find it.
+- Tests, from an agent worktree:
+  `PYTHONPATH=.:packages/cli:packages/mcp:packages/sdk:packages/tui ~/dev/Quorus/.venv/bin/python -m pytest -q -x <test files>`
+- Run only the test files you touched; the laptop is shared, so no full-suite
+  runs unless you're asked to.
+- Lint: `~/dev/Quorus/.venv/bin/ruff check .`

quorus_ai-0.5.0/CHANGELOG.md

@@ -0,0 +1,186 @@
+# Changelog
+
+All notable changes to Quorus are documented here. This project follows
+[Semantic Versioning](https://semver.org/) and [Keep a Changelog](https://keepachangelog.com/).
+
+## [Unreleased]
+
+## [0.5.0] — 2026-10-17
+
+First PyPI release, as `quorus-ai` (`pipx install quorus-ai`; the command is still `quorus`).
+
+### Added
+
+- **Any model.** `quorus agent add opencode --model <provider/model>` runs any model
+  opencode supports: free open-source models (DeepSeek, Qwen), OpenRouter, local Ollama.
+  `quorus agent add gemini` runs Gemini CLI (needs a free AI Studio API key). They get the
+  same room tools, permission modes and session memory as Claude Code and Codex.
+- **As many agents as you want.** `--name backend` adds `<you>-backend-claude` next to
+  `<you>-claude`. `--model` picks the model for any tool.
+- **Every agent answers you.** A human message wakes every agent in the room; agent chatter
+  still picks one.
+- **Team lead.** `quorus agent add claude --room R --lead`: the lead splits your request
+  into tasks and assigns them, the others answer briefly and wait for an assignment,
+  results flow back to the lead, and it tells you when everything is done.
+- **Notifications that reach you.** Your agents ping your desktop when they ship to main,
+  ask you something, or get stuck, even with Quorus closed (`QUORUS_NOTIFY=0` to mute).
+- **See who is working.** `⋯ @you-claude working 1m 20s` in the chat while a wake runs.
+- **Catch up.** The home screen shows what shipped, who needs you and who is stuck since
+  you last opened Quorus.
+- **Readable chat.** Agent markdown (bold, code, bullets, headings) renders as clean text.
+- Warnings before adding an agent that can't wake (no Gemini key, Ollama not running or
+  missing the model).
+
+### Fixed
+
+- opencode agents wrote files into the daemon's working directory instead of the repo, and
+  manual (read-only) rooms weren't read-only on opencode 2.x.
+- A project's own `opencode.json` could point an agent's room tools at the wrong relay.
+- Codex couldn't post to the room itself: its MCP calls needed an approval nobody could give,
+  and a stale `quorus` entry in `~/.codex/config.toml` overrode the agent's identity.
+- Agents committed generated files (`__pycache__`, `.pyc`) to your main branch.
+- Each agent now logs to its own file; a second Ctrl-C no longer prints a traceback.
+- Gemini agents failed on any machine with an old Google login in `~/.gemini`; each Gemini
+  agent now gets its own Gemini home and signs in with your API key.
+- `quorus agent remove` crashed on Linux and `quorus agent list` showed nothing there.
+- opencode 1.x (still what `npm i -g opencode-ai` installs) failed every wake.
+
+## [0.4.0] — 2026-10-10
+
+Never published to PyPI.
+
+### Added
+
+- **Agents that wake on their own.** `quorus agent add claude|codex --room R [--repo P]`
+  starts a background agent that wakes when the room needs it, works in its own git
+  worktree, publishes to your repo's main branch, and asks a teammate to review.
+- **The agent decides whether to answer.** Every message not addressed to someone else
+  wakes one agent, chosen by auction. It answers, or replies `NO_REPLY` to stay quiet.
+  Unaddressed agent chatter stops after 6 agent messages in a row.
+- **Permission modes per room:** `--mode default|manual|autonomous`. Manual asks you
+  before risky actions (`quorus approvals`, `quorus approve <id>`).
+- **First run just works:** `quorus` with no config sets up a local relay with a generated
+  secret, or joins a teammate's relay from an invite.
+- `quorus agent list|remove`, `quorus --version`, readable errors with exit codes
+  (2 unreachable, 3 auth, 4 not found, 5 refused).
+
+### Known limits
+
+- No hosted relay yet: teammates must reach your relay's address.
+- On a shared-secret relay, manual-mode approvals and kicks are advisory. Use account
+  mode for real separation.
+- The background agent uses launchd on macOS and a detached process elsewhere.
+
+### Earlier work included in 0.4.0
+
+### Added — Wave 9 (TUI)
+
+- **Work-queue panel** — `Ctrl-W` toggles a fold-up panel above the chat
+  feed showing every active task in the current room (`task_id`,
+  `claimed_by`, `status`, `started_at`, `eta`). Refetches from
+  `GET /v1/work_queue/{room_id}` on open; cached for 30 s so rapid
+  toggles don't hammer the relay.
+- **Agent-DM inbox** — `Ctrl-N` (originally `Ctrl-D` in the spec, but
+  `Ctrl-D` already binds to QUIT in the TUI input loop and `Ctrl-I`
+  shadows Tab; `Ctrl-N` for "Notifications/iNbox" was free of all
+  collisions) toggles an iMessage-style preview of the last 8 agent
+  DMs. A dedicated background thread subscribes to
+  `/stream/dm/{participant}` and appends events into a bounded deque.
+- **Threaded replies** — when any message in the feed carries
+  `thread_root_id`, `render_bubble_feed` now reorders so children
+  render directly under their root with a `└─` connector and 5-cell
+  left indent. Threads default to **collapsed** (a 1-line
+  `@sender: snippet · N replies   (/expand <id>)` rollup); the new
+  `/expand <root_id|all>` slash command flips an entry to expanded.
+
+Public beta launch. Full rebrand from Murmur, coordination-layer positioning.
+
+### Added
+
+- **Rebrand to Quorus** across package name, CLI entry (`quorus`), config dir
+  (`~/.quorus/`), API key prefix (`quorus_sk_`), join token prefix
+  (`quorus_join_`), and MCP server name.
+- New `quorus-mcp` console entry point — matches the MCP snippet shown on the
+  website so `"command": "quorus-mcp"` works after install.
+- First-class Windsurf integration: `quorus connect windsurf` emits MCP
+  config for `~/.codeium/windsurf/mcp_config.json`.
+- `quorus init --config-dir PATH` flag for multi-profile setups (also honors
+  `$QUORUS_CONFIG_DIR`).
+- Open Graph / Twitter Card social metadata + `og.svg` on the marketing site.
+- `CHANGELOG.md` (this file), `sitemap.xml`, and `robots.txt` for launch
+  hygiene.
+- Copyable one-line install command visible on the website hero.
+- Relay prints a human-friendly listen banner on startup.
+
+### Changed
+
+- **Default `USE_OUTBOX=true`** — fan-out goes through the durable outbox by
+  default; silent message loss on Redis hiccups is gone.
+- Audit service exceptions propagate (per CLAUDE.md Lessons Learned). Callers
+  wrap business + audit writes in a single transaction.
+- `audit_svc.record()` and `record_batch()` accept an optional `session`
+  param so outbox/room writers thread their transaction through — no more
+  nested-session split-brain ledger.
+- Content-Security-Policy tightened: no more `'unsafe-inline'` on
+  `script-src`; added `base-uri`, `form-action`, `img-src data:`.
+- SSE token mint binds `recipient` to the caller identity; legacy auth
+  without admin role can no longer mint tokens for arbitrary recipients.
+- JWT decode errors now log the reason (`ExpiredSignatureError` /
+  `InvalidTokenError`) instead of silently falling through.
+- 404 rate limiter's in-memory dicts are guarded by `asyncio.Lock` — no more
+  read-modify-write races under 404 floods.
+- Doctor separates required vs optional checks (`11/11 required · 0/1
+optional`) so the opt-in hook doesn't read as failure.
+- Website: removed unverified "sub-100ms latency" claim, Copilot logo (no
+  real MCP path), stale "6+ platforms" (now honest 5), and CTA `870+ tests`
+  pill. `quorus-ai` → `quorus` everywhere.
+- `render.yaml` wires JWT_SECRET, BOOTSTRAP_SECRET, DATABASE_URL, REDIS_URL,
+  USE_OUTBOX, ALLOW_LEGACY_AUTH, CORS_ORIGINS.
+
+### Fixed
+
+- `/v1/usage` 500 crash for non-admins (`list_by_member` → `list_for_member`).
+- Cursor + Gemini `quorus connect` commands emitted MCP configs pointing at a
+  nonexistent path; now emit `"command": "quorus-mcp"`.
+- `quorus join <quorus_join_...>` previously rejected new-prefix tokens; now
+  accepts both `quorus_join_` and legacy `murm_join_`.
+- Config file TOCTOU race — now `os.open(..., 0o600)` atomic write.
+- Hero particle canvas respects `prefers-reduced-motion` and pauses when
+  off-screen (battery + INP).
+- Waitlist emails: rebrand sender defaults to `@quorus.dev`, use `waitUntil`
+  so Vercel serverless doesn't freeze before send completes.
+- Footer anchor `#howit` pointed at a non-rendered section; now `#architecture`.
+- AgentShowcase branding banner: MURMUR → QUORUS.
+- TUI guard against `selected is None` crash on fresh install.
+- Vite sourcemaps no longer shipped as public `.map` files.
+
+### Security
+
+- See `Changed` above for CSP, SSE-binding, JWT logging, and 404 race fixes.
+- Four critical items fixed this cycle: osascript injection in watcher,
+  agent-name shell injection in spawn, bootstrap-secret timing attack
+  (`hmac.compare_digest`), config-write TOCTOU.
+- IP rate limits added to invite-join (10/min) and admin tenant creation
+  (5/min).
+- Invite page template values are `html.escape`d.
+
+### Known limitations (beta-framed)
+
+- Not on PyPI yet — install via `pipx install "quorus @ git+..."`.
+- Load/chaos tests not yet run. Concurrency envelope unknown.
+- Webhook egress relies on app-level private-IP blocklist (no network-level
+  proxy).
+- Admin tooling for DLQ replay and tenant suspension is CLI-only.
+
+---
+
+## [0.3.0] — 2026-04-13 (pre-rebrand, Murmur)
+
+- Monorepo split into packages/sdk, packages/cli, packages/mcp, packages/tui.
+- Transactional outbox pattern (optional, `USE_OUTBOX=true`).
+- Audit ledger with `/v1/audit/*` endpoints.
+- Account-based identity — `participant_id` in JWT claims + FK columns.
+
+## [0.2.x] — earlier (pre-outbox)
+
+See git history.

quorus_ai-0.5.0/CLAUDE.md

@@ -0,0 +1,59 @@
+# CLAUDE.md — Quorus
+
+## Shared Context Protocol
+
+**Read `docs/internal/CONTEXT.md` at the start of every session.** It contains the current project state, what's in progress, recent changes, architectural decisions, and next priorities. It is the shared memory between all contributors' Claude instances.
+
+**Update `docs/internal/CONTEXT.md` after every significant change:**
+
+- Add your changes to "Recent Changes" (keep last 10, drop oldest)
+- Update "In Progress" when starting/finishing work
+- Update "Current State" if the system's capabilities changed
+- Add to "Decisions Made" when architectural choices are made
+- Update "Next Up" when priorities shift
+
+---
+
+## Project
+
+Quorus — coordination layer for AI agent swarms. Real-time rooms, messaging, distributed locks.
+
+**Stack:** Python 3.10+, FastAPI, asyncio, httpx, mcp (FastMCP), pytest, ruff, rich
+
+**Run tests:** `pytest -v`
+**From an agent worktree** (no venv of its own): `PYTHONPATH=.:packages/cli:packages/mcp:packages/sdk:packages/tui ~/dev/Quorus/.venv/bin/python -m pytest -q`
+**Lint:** `ruff check .`
+**Lint fix:** `ruff check . --fix`
+
+---
+
+## Architecture
+
+```
+Claude Code A -> MCP Server (stdio) -> HTTP -> [Quorus Relay (FastAPI)] -> HTTP -> MCP Server (stdio) -> Claude Code B
+```
+
+- `quorus/relay.py` — Central relay (FastAPI). Rooms, SSE fan-out, persistence, rate limiting.
+- `quorus_mcp/server.py` — MCP server. 12 tools for coordination.
+- `quorus/config.py` — Config loading. Priority: env vars > ~/.quorus/config.json > legacy fallback.
+- `quorus_cli/cli.py` — CLI commands. `quorus` opens TUI by default.
+
+---
+
+## Rules
+
+- Files under 500 lines. Split if larger.
+- Async-first. Use `asyncio.to_thread` for blocking I/O.
+- All external input validated before use.
+- Never log secrets or tokens.
+- Tests required for new features and bug fixes.
+- Conventional commits, under 50 chars, imperative mood.
+
+## graphify
+
+This project has a graphify knowledge graph at graphify-out/.
+
+Rules:
+- Before answering architecture or codebase questions, read graphify-out/GRAPH_REPORT.md for god nodes and community structure
+- If graphify-out/wiki/index.md exists, navigate it instead of reading raw files
+- After modifying code files in this session, run `graphify update .` to keep the graph current (AST-only, no API cost)

quorus_ai-0.5.0/CONTRIBUTING.md

@@ -0,0 +1,58 @@
+# Contributing to Murmur
+
+Thanks for your interest in contributing to Murmur!
+
+## Development Setup
+
+```bash
+git clone https://github.com/Quorus-dev/Quorus.git
+cd Quorus
+python3 -m venv .venv && source .venv/bin/activate
+pip install -e ".[dev]" -e packages/sdk -e packages/cli -e packages/mcp -e packages/tui
+# macOS only: undo a hatchling-editable + Python 3.14 + macOS interaction
+# that hides .pth files via com.apple.provenance + UF_HIDDEN. Idempotent.
+bash scripts/fix_editable_pth.sh
+python -c "import quorus, quorus_sdk, quorus_cli, quorus_mcp, quorus_tui"  # must print nothing
+```
+
+## Running Tests
+
+```bash
+pytest -v
+```
+
+## Running the E2E Test
+
+```bash
+python scripts/e2e_live.py
+```
+
+## Code Standards
+
+- Python 3.10+, async-first
+- Ruff for linting: `ruff check .`
+- Max 100 char lines, files under 500 lines
+- All external input validated before use
+- Never log secrets or tokens
+- Tests required for new features and bug fixes
+
+## Commit Conventions
+
+- Conventional commits: `feat:`, `fix:`, `refactor:`, `test:`, `docs:`, `chore:`
+- Under 50 chars, imperative mood
+- One logical change per commit
+
+## Submitting Changes
+
+1. Fork the repo and create a branch
+2. Make your changes with tests
+3. Run `ruff check .` and `pytest -v`
+4. Submit a pull request
+
+## Architecture
+
+See `docs/internal/CONTEXT.md` for current architecture and design decisions.
+
+## License
+
+By contributing, you agree that your contributions will be licensed under the MIT License.

quorus_ai-0.5.0/Dockerfile

@@ -0,0 +1,70 @@
+# syntax=docker/dockerfile:1.7
+# ---------------------------------------------------------------------------
+# Stage 1 — builder: install pinned dependencies + build the wheel.
+# Ships pip + build deps + cached wheels.
+# ---------------------------------------------------------------------------
+FROM python:3.12-slim AS builder
+
+WORKDIR /build
+
+# Build deps + uv (10x faster pip resolver, deterministic locks).
+ENV PIP_NO_CACHE_DIR=1 \
+    PIP_DISABLE_PIP_VERSION_CHECK=1 \
+    PYTHONDONTWRITEBYTECODE=1
+
+# Install into an isolated prefix so we can copy the resolved tree forward
+# into the runtime stage without dragging pip / wheels along with it.
+ENV PREFIX=/install
+RUN mkdir -p "$PREFIX"
+
+# Pinned deps from the lockfile (single source of truth — see pyproject.toml
+# header). The lockfile must be regenerated via ``uv pip compile`` whenever
+# pyproject.toml changes. Both files commit in the same PR.
+COPY requirements.lock ./
+RUN pip install --no-cache-dir --prefix="$PREFIX" -r requirements.lock
+
+# Source layout: hatchling's wheel pulls every package listed in
+# [tool.hatch.build.targets.wheel].packages — quorus/ + packages/* monorepo.
+COPY pyproject.toml README.md alembic.ini ./
+COPY quorus/ quorus/
+COPY packages/ packages/
+
+# Install Quorus itself (no deps — those came from the lockfile above).
+RUN pip install --no-cache-dir --prefix="$PREFIX" --no-deps .
+
+
+# ---------------------------------------------------------------------------
+# Stage 2 — runtime: lean image, no pip / build deps / wheel cache.
+# ---------------------------------------------------------------------------
+FROM python:3.12-slim AS runtime
+
+WORKDIR /app
+
+# Carry the resolved Python install forward; pip itself is left behind.
+COPY --from=builder /install /usr/local
+# Fly release_command runs from /app, so Alembic's config and migration
+# scripts must exist outside the installed wheel.
+COPY --from=builder /build/alembic.ini /app/alembic.ini
+COPY --from=builder /build/quorus/migrations /app/quorus/migrations
+
+# Non-root user for security. Match the data dir's UID so volumes mounted
+# with the default UID:GID (1000:1000) work without a bind-mount chown.
+RUN useradd --create-home --shell /bin/bash --uid 1000 quorus \
+    && mkdir -p /app/data \
+    && chown -R quorus:quorus /app/data
+
+USER quorus
+
+# Persist messages across restarts (fallback when Postgres unavailable).
+VOLUME ["/app/data"]
+ENV MESSAGES_FILE=/app/data/messages.json
+
+EXPOSE 8080
+ENV PORT=8080 \
+    PYTHONDONTWRITEBYTECODE=1 \
+    PYTHONUNBUFFERED=1
+
+HEALTHCHECK --interval=30s --timeout=5s --retries=3 \
+  CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:8080/health')"
+
+CMD python -m uvicorn quorus.relay:app --host 0.0.0.0 --port ${PORT:-8080}

quorus_ai-0.5.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Arav Kekane, Aarya
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.