atlas-os 0.3.0__tar.gz

atlas_os-0.3.0/.env.example

@@ -0,0 +1,97 @@
+# ──────────────────────────────────────────────────────────────────────────────
+# Atlas OS configuration
+#
+# Copy this file to `.env` and fill in your own values:
+#     cp .env.example .env
+#
+# Load it into your shell before running scripts:
+#     set -a; source .env; set +a
+#
+# `.env` is git-ignored. NEVER commit real secrets. All values below are
+# placeholders / safe defaults — none are real credentials.
+# ──────────────────────────────────────────────────────────────────────────────
+
+# ── Vault ───────────────────────────────────────────────────────────────────
+# Absolute path to your markdown vault. Required by almost every script.
+VAULT_PATH=~/Documents/Obsidian/MyVault
+
+# Where to write the RAG vector store and graph (default: $VAULT_PATH/.rag)
+# RAG_DIR=~/Documents/Obsidian/MyVault/.rag
+
+# Where your Claude scheduled-task SKILL.md folders live
+# SCHEDULED_DIR=~/Documents/Claude/Scheduled
+
+# Where `atlas skills install <name>` writes installed skills
+# (default: $VAULT_PATH/.claude/skills)
+# ATLAS_SKILLS_DIR=~/Documents/Claude/Scheduled
+
+# ── Pluggable LLM backends (auto-detected) ───────────────────────────────────
+# Atlas OS talks to any OpenAI-compatible LLM server. If you leave the EMBED_*
+# and LM_STUDIO_* vars below unset, it auto-detects a running backend by probing
+# (in order): LM Studio → Ollama → llama.cpp → a custom OpenAI-compatible URL.
+# Run `atlas backends` to see what's detected, `atlas backends test` to verify.
+#
+#   ATLAS_LLM_BACKEND   force one backend, skipping detection:
+#                       lmstudio | ollama | llamacpp | openai-compatible
+#   ATLAS_LLM_MODEL     override the chat model name reported to callers
+# ATLAS_LLM_BACKEND=
+# ATLAS_LLM_MODEL=
+#
+# Per-backend base URL overrides (only needed if not on the default localhost port):
+#   LM_STUDIO_URL (default :5555) · OLLAMA_URL (default :11434)
+#   LLAMACPP_URL (default :8080)  · OPENAI_COMPATIBLE_URL (no default — must be set)
+# OLLAMA_URL=http://localhost:11434
+# LLAMACPP_URL=http://localhost:8080
+# OPENAI_COMPATIBLE_URL=
+
+# ── Local LLM: embeddings (OpenAI-compatible, e.g. LM Studio / Ollama) ───────
+# Explicit values here take precedence over backend auto-detection above.
+EMBED_HOST=localhost
+EMBED_PORT=5555
+EMBED_MODEL=text-embedding-nomic-embed-text-v1.5
+# EMBED_API_KEY=                 # only if your endpoint requires a bearer token
+# EMBED_URL=                     # overrides host/port if set
+
+# ── Local LLM: chat completions (used by trading module) ─────────────────────
+LM_STUDIO_HOST=localhost
+LM_STUDIO_PORT=5555
+LM_STUDIO_MODEL=local-model
+# Two scripts read the chat endpoint and expect DIFFERENT shapes — set whichever
+# you use, or leave both unset to fall back to host/port:
+#   LM_STUDIO_URL       used by scripts/trading_briefing.py — include the /v1 suffix
+#                       default: http://$LM_STUDIO_HOST:$LM_STUDIO_PORT/v1
+#   LM_STUDIO_ENDPOINT  used by trading/config.py + trading/core.py — NO /v1 suffix
+#                       default: http://$LM_STUDIO_HOST:$LM_STUDIO_PORT
+# LM_STUDIO_URL=
+# LM_STUDIO_ENDPOINT=
+
+# ── Text-to-speech (optional) ────────────────────────────────────────────────
+TTS_HOST=localhost
+TTS_PORT=8800
+
+# ── Email (SMTP) ──────────────────────────────────────────────────────────────
+# The account that sends reports/newsletters.
+SENDER_EMAIL=your-atlas-account@example.com
+SENDER_NAME=Atlas
+SMTP_SERVER=smtp.gmail.com
+SMTP_PORT=587
+# Gmail app password (https://myaccount.google.com/apppasswords) — NEVER commit
+SMTP_APP_PASSWORD=your-app-password
+
+# Where reports are sent (usually yourself)
+USER_EMAIL=your-email@example.com
+
+# ── Dashboard ports (optional) ────────────────────────────────────────────────
+DASHBOARD_FRONTEND_PORT=3000
+DASHBOARD_BACKEND_PORT=5001
+
+# ── Trading module (optional) ─────────────────────────────────────────────────
+TRADING_AGENTS_PATH=~/Documents/TradingAgents
+TRADING_TICKERS=BTC-USD,ETH-USD
+# Cloud Portfolio Manager step (off by default) — only if you opt in:
+# ANTHROPIC_API_KEY=sk-ant-your-key
+# ANTHROPIC_MODEL=claude-opus-4-6
+
+# ── Git (optional) ────────────────────────────────────────────────────────────
+# If you mirror your vault to a remote, set it here (your private repo).
+# GITHUB_REPO=your-username/your-vault

atlas_os-0.3.0/.gitignore

@@ -0,0 +1,76 @@
+# ── Secrets & credentials ───────────────────────────────────────────────────
+.env
+.env.*
+!.env.example
+*.key
+*.pem
+*.p12
+credentials*
+*credential*
+*password*
+*secret*
+*.token
+.anthropic-api-key
+*-app-password
+*gmail-app-password*
+
+# ── Personal data / PII (never commit) ──────────────────────────────────────
+*.xlsx          # job trackers and other spreadsheets contain PII
+*.xls
+*.csv           # may contain personal/exported data
+vault-content/
+vault/
+personal/
+private/
+job-search/
+job-tracker*
+briefings/
+*briefing*.md
+session-logs/
+*.eml
+
+# ── Generated RAG / graph data (contains embedded personal notes) ───────────
+*.sqlite
+*.sqlite3
+*.db
+vectors.json
+vectors.json.tmp
+graph.json
+graph.json.tmp
+last_embed.txt
+last_embed_fallback.txt
+embed_checkpoint.json
+.rag/
+.vectors/
+cache/
+
+# ── Trading (local config may contain real paths / keys) ────────────────────
+trading/config.py
+signals.json
+
+# ── Python ──────────────────────────────────────────────────────────────────
+__pycache__/
+*.py[cod]
+*.egg-info/
+.venv/
+venv/
+.env.bak
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+
+# ── Node / dashboard ────────────────────────────────────────────────────────
+node_modules/
+dist/
+build/
+*.log
+
+# ── OS / editor ─────────────────────────────────────────────────────────────
+.DS_Store
+Thumbs.db
+*.swp
+.idea/
+.vscode/
+
+# ── Local Claude Code / agent tooling state (not part of the public repo) ────
+.claude/

atlas_os-0.3.0/CHANGELOG.md

@@ -0,0 +1,249 @@
+# Changelog
+
+All notable changes to Atlas OS are documented here. The format is based on
+[Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and this project
+aims to follow [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+### Added
+- **Automated PyPI publishing via GitHub Actions + Trusted Publishing.** Pushing
+  a `v*` tag (e.g. `v0.4.0`) now builds, tests, and publishes the release to
+  PyPI with no stored token — authentication is
+  [OIDC Trusted Publishing](https://docs.pypi.org/trusted-publishers/).
+  [`.github/workflows/publish.yml`](.github/workflows/publish.yml) runs
+  `test → build → publish`: lint + `pytest`, then `python -m build` +
+  `twine check`, then `pypa/gh-action-pypi-publish` against the `pypi`
+  environment with `id-token: write`. A companion
+  [`.github/workflows/test-publish.yml`](.github/workflows/test-publish.yml)
+  routes pre-release tags (`v*rc*`, `v*dev*`) to TestPyPI for dress rehearsals.
+  [`docs/PUBLISHING.md`](docs/PUBLISHING.md) documents the flow and the one-time
+  PyPI trusted-publisher setup (pending-publisher or manual-first-upload).
+- **`atlas skills install <name>`** — one-command skill deployment. Copies a
+  skill's `SKILL.md` into your scheduled-tasks directory (`$ATLAS_SKILLS_DIR`,
+  default `$VAULT_PATH/.claude/skills/<name>/`) and substitutes its
+  `{{PLACEHOLDER}}` tokens from the environment / `.env`. Most tokens map to the
+  env var of the same name (`{{VAULT_PATH}}` ← `VAULT_PATH`); `{{ATLAS_OS}}`
+  resolves to the repo path and `{{LLM_PORT}}` reads `LM_STUDIO_PORT`. Tokens
+  with no value are left in place and reported so you can fill them by hand;
+  `--force` overwrites an existing install. Two companion subcommands:
+  `atlas skills list` (every available skill) and `atlas skills show <name>`
+  (print a skill's `SKILL.md`). The placeholder logic and install live in
+  [`atlas_os/_skills.py`](atlas_os/_skills.py), covered by
+  [`tests/test_skills.py`](tests/test_skills.py).
+- **PyPI release preparation.** The package is now publish-ready: the version is
+  single-sourced from `__version__` in
+  [`atlas_os/__init__.py`](atlas_os/__init__.py) via `[tool.hatch.version]`
+  (bump it in one place), the sdist explicitly bundles `scripts/`, `schemas/`,
+  `templates/`, `skills/`, and `docs/` (`[tool.hatch.build.targets.sdist]`), and
+  the metadata gained AI/indexing topic classifiers. A new
+  [`docs/PUBLISHING.md`](docs/PUBLISHING.md) is the maintainer runbook (bump →
+  `python -m build` → `twine check` → TestPyPI → `twine upload` → tag). No
+  release is published yet — everything is staged so `pipx install atlas-os`
+  works the moment it is.
+- **Pluggable LLM backends** ([`atlas_os/backends.py`](atlas_os/backends.py)).
+  Atlas OS now auto-detects any OpenAI-compatible LLM server, probing **LM Studio
+  (`:5555`) → Ollama (`:11434`) → llama.cpp (`:8080`) → a custom
+  `OPENAI_COMPATIBLE_URL`** in that order and using the first that responds. The
+  module exposes `detect_backend()`, `get_client()`, `list_models()`, and a
+  one-shot `run_inference()` test. Two env vars override the defaults:
+  `ATLAS_LLM_BACKEND` forces a backend (`lmstudio`/`ollama`/`llamacpp`/
+  `openai-compatible`, skipping detection) and `ATLAS_LLM_MODEL` overrides the
+  chat model name. New CLI: `atlas backends` lists every backend with
+  reachability + models, and `atlas backends test` runs an end-to-end inference.
+  The RAG (`embed_vault.py`), trading (`trading_briefing.py`), and health
+  (`health_check.py`) scripts now resolve their endpoint through this module.
+  **Fully backward compatible:** explicit `EMBED_*` / `LM_STUDIO_*` settings still
+  win, so existing setups are unchanged. Covered by
+  [`tests/test_backends.py`](tests/test_backends.py).
+- **Skills catalogue & framework docs.** A
+  [`docs/SKILLS-CATALOGUE.md`](docs/SKILLS-CATALOGUE.md) documenting the full menu
+  of **160+ skills** — 149 capability skills across seven domains (Security,
+  DevOps, Frontend, Backend, Quality, Data & AI, Business), each with a one-line
+  summary, what it does, and when to use it, plus the four Atlas-native skills
+  (`autoresearch`, `save-to-vault`, `wiki-search`, `send-email`) and the nine
+  scheduled automations. A companion
+  [`docs/SKILLS-FRAMEWORK.md`](docs/SKILLS-FRAMEWORK.md) explains what a skill is,
+  the anatomy of a `SKILL.md`, the placeholder-token system, the lifecycle
+  (creation → installation → scheduling → execution → audit logging), how the
+  RAG-indexed catalog reaches sub-agents, the `skill-creator` meta-skill, and a
+  copy-paste `SKILL.md` template. Everything is generic and `{{PLACEHOLDER}}`-
+  tokenised — no personal data.
+- **Six example skills** ([`skills/`](skills)) demonstrating the framework across
+  document creation (`generate-vault-report-doc`), email automation
+  (`inbox-triage-digest`), data analysis (`spreadsheet-analysis`), web research
+  (`topic-research-brief`), vault management (`vault-lint-report`), and report
+  generation (`weekly-digest-report`) — bringing the shipped skill count to 15.
+- **Audit trail** ([`atlas_os/audit.py`](atlas_os/audit.py)). An append-only
+  JSONL log of every autonomous action, written to `$ATLAS_AUDIT_PATH` (default
+  `$VAULT_PATH/.atlas/audit.jsonl`). Each entry records the timestamp, action,
+  trigger (`scheduled`/`manual`/`cli`), status, duration, what changed, why it
+  ran, and any error. Appends are serialised with an in-process lock plus an
+  OS-level advisory file lock (safe across concurrent `atlas` processes) and the
+  file auto-rotates at 10 MB to `audit.jsonl.1`, `.2`, …. Every script-wrapping
+  command (`embed`, `commit`, `graph`, `changelog`, `health`, `trading`,
+  `email`) now logs its outcome automatically; scheduled tasks tag their runs by
+  setting `ATLAS_TRIGGER=scheduled`. New `atlas audit` subcommands: `show`
+  (`--limit`/`--action`/`--since`), `tail` (last 5, compact), and `export`
+  (`--format csv|json`, `--output`) for compliance reporting. Strengthens ISO
+  27001 control A.12.4 (logging & monitoring). Covered by
+  [`tests/test_audit.py`](tests/test_audit.py).
+- **GitHub issue & PR templates**
+  ([`.github/ISSUE_TEMPLATE/`](.github/ISSUE_TEMPLATE/),
+  [`.github/PULL_REQUEST_TEMPLATE.md`](.github/PULL_REQUEST_TEMPLATE.md)) — a
+  structured bug report (with an environment block: OS, Python, Atlas OS
+  version, install method), a feature request (use case / proposed solution /
+  alternatives), and a PR checklist (tests, docs, the PII scan). An
+  `ISSUE_TEMPLATE/config.yml` disables blank issues and routes security reports
+  to `SECURITY.md` and questions to the FAQ.
+- **README status badges** — GitHub Actions CI status, GitHub stars, and
+  last-commit, alongside the existing license / Python / privacy / docs badges.
+- **`atlas trading`** — wraps `scripts/trading_briefing.py` (`--ticker`,
+  `--date`, `--dry-run`); the last optional pipeline script to gain a first-class
+  subcommand, so the whole system is now reachable through one `atlas` command.
+- **Up-front env validation.** Every vault/optional command checks its required
+  environment variables before shelling out and exits with a clear message and a
+  non-zero code if any are missing — a half-configured feature fails fast instead
+  of part-way through.
+- **`atlas email` flags.** Send mail with `--to` / `--subject` / `--body`
+  (`--text` for plain text, repeatable `--attach`), or the original raw payload
+  via `--json`.
+- **Docker support.** A minimal [`Dockerfile`](Dockerfile) (Python 3.11-slim +
+  git) that packages the `atlas` CLI, a [`docker-compose.yml`](docker-compose.yml)
+  that bind-mounts your vault and loads `.env`, and a `.dockerignore`. Run any
+  subcommand in a container without installing Python tooling on the host.
+  Build-tested end-to-end (`atlas --version` / `doctor` / `commit` against a
+  bind-mounted vault), with three fixes from that pass: copy the
+  `scripts/schemas/templates/skills` dirs *before* `pip install` (the wheel
+  force-includes them, so the build failed without them); `git config --global
+  --add safe.directory` so git operations work on a vault owned by a non-root
+  host user (avoids "dubious ownership"); and an optional `env_file` so compose
+  runs before a `.env` exists.
+- **CLI tests** ([`tests/test_cli.py`](tests/test_cli.py)) covering `--version`,
+  every registered subcommand, and the env-validation guards.
+- **Core vs optional** section in [`docs/SETUP.md`](docs/SETUP.md) separating the
+  always-available core (vault, commit, changelog, schemas, health) from opt-in
+  features (RAG/embeddings, trading, email, LM Studio, dashboard) with each one's
+  extra deps and env vars, plus a Docker quick-start.
+- **Automated test suite** in [`tests/`](tests/) — 74 hermetic `pytest` tests
+  covering the core scripts (`embed_vault`, `build_graph`, `health_check`,
+  `send_email`, `vault_commit`, `vault_changelog`, `trading_briefing`). They stub
+  every external dependency (network, SMTP, git, and the optional
+  `tradingagents` package) and point `VAULT_PATH`/`RAG_DIR` at a temp directory,
+  so they need no env vars, no network, and never touch a real vault.
+- **GitHub Actions CI** ([`.github/workflows/ci.yml`](.github/workflows/ci.yml))
+  — runs ruff, pytest, and pip-audit on every push and pull request to `main`.
+- **Development & testing** section in the README and dev/CI tooling (`pytest`,
+  `ruff`, `pip-audit`) added to `requirements.txt`.
+- **Feature deep-dive docs** in [`docs/features/`](docs/features/README.md) — one
+  per feature, explaining how it actually works (internals, data formats,
+  configuration, edge cases), grounded in the source: knowledge vault & schemas,
+  local RAG search, knowledge graph, git automation, scheduled tasks & skills
+  catalog, email reports, trading SDK, and health check & dashboard. Linked from
+  the README and the docs index.
+
+### Changed
+- **`CONTRIBUTING.md` expanded** into a full contributor guide — dev-environment
+  setup, running the test/lint/audit suite, code style, the PR workflow, and a
+  project-structure overview — on top of the existing "golden rule" (never
+  commit personal data) and PII scan.
+- **README**: documented `atlas trading`, the new `atlas email` flags,
+  env-validation behaviour, a Docker section (+ a Docker pointer under
+  Installation), the Docker files in the repo-layout diagram, and updated the
+  `.github/` layout line to note the issue/PR templates.
+- **Docs sync across the set** to match the streamlined CLI:
+  - `docs/ARCHITECTURE.md` — added the `atlas` CLI as component 0 (the unified
+    entry point) and a Deployment section covering checkout / installed tool /
+    Docker.
+  - `docs/EXAMPLES.md`, `docs/features/email-reports.md` — updated every
+    `atlas email` example to the new flags (with `--json` for raw payloads),
+    fixing samples that the flag change would otherwise have broken.
+  - `docs/QUICKSTART.md` — corrected the minimum Python to 3.11+ and linked the
+    Docker quick-start.
+  - `docs/README.md` (docs index) — linked the SETUP core-vs-optional matrix and
+    the Docker files.
+  - `SECURITY.md` — added a "Running in containers (Docker)" section (no secrets
+    or vault data in the image; runtime-only `--env-file`; bind-mounted vault).
+- **`pyproject.toml`** continues to declare the `atlas` entry point and the
+  optional dependency groups (`[pdf]`, `[trading]`, `[all]`); these are now the
+  documented install path (`pip install -e ".[all]"`) for the optional features.
+
+## [0.3.0] — 2026-06-02
+
+### Added
+- **Agent skills catalog.** A self-updating `Skills Catalog.md` note generated
+  into the vault, listing every skill (name, description, suggested cadence)
+  read from each `skills/*/SKILL.md` frontmatter — so agents reading or
+  searching the vault can discover what automations they can invoke. Carries
+  `type: reference` frontmatter so the RAG indexer picks it up.
+- **`atlas skills`** — list the catalog in the terminal; `atlas skills --sync`
+  (re)generates the note in the vault (`--output` to override the path).
+  `atlas init` now generates it automatically on setup.
+- The `skills/` directory is bundled into the wheel so the catalog works in an
+  installed `atlas` without the source checkout.
+
+## [0.2.0] — 2026-06-02
+
+### Added
+- **Installable package** with a unified **`atlas` CLI** (`pyproject.toml`,
+  `atlas_os/`). Install via `uv tool install` / `pipx` / `pip install -e .`.
+  Subcommands `embed`, `graph`, `commit`, `changelog`, `health`, `email`, and
+  `schemas` wrap the existing scripts and forward their flags; configuration
+  (`.env`) is auto-loaded.
+- **`atlas init`** — guided onboarding: auto-detects a local LLM (LM Studio /
+  Ollama / any OpenAI-compatible endpoint), writes `.env`, scaffolds the vault
+  skeleton, initialises the vault git repo, and optionally installs the
+  `CLAUDE.md` template. Supports `--vault`, `--yes`, `--force`.
+- **`atlas doctor`** — validates the whole setup (Python, vault, git, RAG
+  index, embeddings endpoint, SMTP) and reports OK / WARN / FAIL with a
+  non-zero exit on failures.
+- Optional dependency extras: `atlas-os[trading]` (yfinance),
+  `atlas-os[pdf]` (pdfplumber), `atlas-os[all]`.
+- `docs/CONFIGURATION.md` — authoritative reference for every environment
+  variable (purpose, default, required/optional, consuming scripts).
+- `docs/SCRIPTS.md` — complete CLI reference for all scripts and their flags,
+  including the previously-undocumented `embed_vault.py` flags
+  (`--incremental`, `--folder`, `--pdfs-only`, `--checkpoint-interval`,
+  `--batch-size`).
+- `docs/FAQ.md` — frequently asked questions and troubleshooting.
+- `docs/README.md` — documentation index and recommended reading order.
+- `CHANGELOG.md` — this file.
+
+### Changed
+- The scripts/schemas/templates are now bundled into the wheel (under
+  `atlas_os_data/`) so an installed `atlas` works without the source checkout;
+  in a source checkout the CLI uses the live files.
+- `.env.example` now documents `LM_STUDIO_URL` (used by
+  `scripts/trading_briefing.py`, expects a `/v1` suffix) alongside
+  `LM_STUDIO_ENDPOINT` (used by `trading/config.py`/`core.py`, no suffix),
+  clarifying which script reads which.
+- Root `README.md` expanded with the install/CLI quick start, a CLI command
+  table, badges, a documentation map, and a configuration pointer.
+- `docs/SETUP.md` restructured into "install the package" (recommended) vs
+  "run from a source checkout".
+
+## [0.1.0] — 2026-06-02
+
+### Added
+- Initial public release of Atlas OS — a local-first personal AI operating
+  system built on Claude Cowork.
+- Knowledge vault conventions with per-folder frontmatter schema enforcement
+  (`schemas/`).
+- Local RAG pipeline (`scripts/embed_vault.py`) and wikilink knowledge graph
+  (`scripts/build_graph.py`).
+- Git automation: auto-commit (`scripts/vault_commit.py`) and changelog
+  (`scripts/vault_changelog.py`).
+- Credential-free SMTP email sender (`scripts/send_email.py`).
+- System health check across all subsystems (`scripts/health_check.py`).
+- Optional multi-agent trading research SDK (`trading/`) and briefing generator
+  (`scripts/trading_briefing.py`).
+- Nine Claude Cowork scheduled-task skills (`skills/`).
+- Templates for `CLAUDE.md`, memory structure, vault skeleton, and a static ops
+  dashboard (`templates/`, `dashboard/`).
+- Documentation: setup, architecture, rebuild runbook, scheduled tasks, data
+  classification; `SECURITY.md`, `CONTRIBUTING.md`, MIT `LICENSE`.
+
+[Unreleased]: https://github.com/paulholland511/atlas-os/compare/v0.3.0...HEAD
+[0.3.0]: https://github.com/paulholland511/atlas-os/compare/v0.2.0...v0.3.0
+[0.2.0]: https://github.com/paulholland511/atlas-os/compare/v0.1.0...v0.2.0
+[0.1.0]: https://github.com/paulholland511/atlas-os/releases/tag/v0.1.0

atlas_os-0.3.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Atlas OS contributors
+
+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.