agent-os-kernel 1.1.0__tar.gz

agent_os_kernel-1.1.0/.gitignore

@@ -0,0 +1,154 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+Pipfile.lock
+
+# PEP 582
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# IDE
+.idea/
+.vscode/
+*.swp
+*.swo
+*~
+
+# OS
+.DS_Store
+Thumbs.db
+
+# Project specific
+*.pypirc
+.env.local
+.env.*.local
+
+# Benchmark results (keep templates, ignore generated)
+*_results.json
+benchmark_results/
+
+# Experiment outputs
+experiments/outputs/
+experiments/logs/

agent_os_kernel-1.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024-2026 Imran Siddique
+
+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.

agent_os_kernel-1.1.0/PKG-INFO

@@ -0,0 +1,400 @@
+Metadata-Version: 2.4
+Name: agent-os-kernel
+Version: 1.1.0
+Summary: A Safety-First Kernel for Autonomous AI Agents - POSIX-inspired primitives with 0% policy violation guarantee
+Project-URL: Homepage, https://github.com/imran-siddique/agent-os
+Project-URL: Documentation, https://github.com/imran-siddique/agent-os/tree/master/docs
+Project-URL: Repository, https://github.com/imran-siddique/agent-os
+Project-URL: Bug Tracker, https://github.com/imran-siddique/agent-os/issues
+Project-URL: Changelog, https://github.com/imran-siddique/agent-os/blob/master/CHANGELOG.md
+Author-email: Imran Siddique <imran.siddique@microsoft.com>
+Maintainer-email: Imran Siddique <imran.siddique@microsoft.com>
+License-Expression: MIT
+License-File: LICENSE
+Keywords: agents,ai,autonomous,control-plane,governance,kernel,llm,multi-agent,posix,safety,signals,vfs
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Security
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: System :: Operating System Kernels
+Classifier: Typing :: Typed
+Requires-Python: >=3.9
+Requires-Dist: pydantic>=2.0.0
+Provides-Extra: all
+Requires-Dist: agent-os[dev,docs,full]; extra == 'all'
+Provides-Extra: amb
+Requires-Dist: aiofiles>=23.0.0; extra == 'amb'
+Requires-Dist: anyio>=3.0.0; extra == 'amb'
+Provides-Extra: atr
+Requires-Dist: docker>=6.0.0; extra == 'atr'
+Provides-Extra: caas
+Requires-Dist: fastapi>=0.100.0; extra == 'caas'
+Requires-Dist: numpy>=1.20.0; extra == 'caas'
+Requires-Dist: pypdf2>=3.0.0; extra == 'caas'
+Requires-Dist: scikit-learn>=1.0.0; extra == 'caas'
+Requires-Dist: uvicorn>=0.20.0; extra == 'caas'
+Provides-Extra: cmvk
+Requires-Dist: numpy>=1.20.0; extra == 'cmvk'
+Provides-Extra: control-plane
+Provides-Extra: core
+Requires-Dist: agent-os[cmvk,control-plane,primitives]; extra == 'core'
+Provides-Extra: dev
+Requires-Dist: black>=23.0.0; extra == 'dev'
+Requires-Dist: mypy>=1.0.0; extra == 'dev'
+Requires-Dist: pytest-asyncio>=0.21.0; extra == 'dev'
+Requires-Dist: pytest-cov>=4.0.0; extra == 'dev'
+Requires-Dist: pytest>=7.0.0; extra == 'dev'
+Requires-Dist: ruff>=0.1.0; extra == 'dev'
+Provides-Extra: docs
+Requires-Dist: mkdocs-material>=9.0.0; extra == 'docs'
+Requires-Dist: mkdocs>=1.5.0; extra == 'docs'
+Requires-Dist: mkdocstrings[python]>=0.24.0; extra == 'docs'
+Provides-Extra: emk
+Requires-Dist: chromadb>=0.4.0; extra == 'emk'
+Requires-Dist: numpy>=1.20.0; extra == 'emk'
+Provides-Extra: full
+Requires-Dist: agent-os[amb,atr,caas,cmvk,control-plane,emk,iatp,mute-agent,primitives,scak]; extra == 'full'
+Provides-Extra: iatp
+Requires-Dist: fastapi>=0.100.0; extra == 'iatp'
+Requires-Dist: httpx>=0.24.0; extra == 'iatp'
+Requires-Dist: uvicorn>=0.20.0; extra == 'iatp'
+Provides-Extra: infrastructure
+Requires-Dist: agent-os[amb,atr,iatp]; extra == 'infrastructure'
+Provides-Extra: intelligence
+Requires-Dist: agent-os[mute-agent,scak]; extra == 'intelligence'
+Provides-Extra: mute-agent
+Provides-Extra: primitives
+Provides-Extra: scak
+Requires-Dist: pyyaml>=6.0; extra == 'scak'
+Description-Content-Type: text/markdown
+
+<div align="center">
+
+# Agent OS
+
+### The Linux Kernel for AI Agents
+
+**0% Safety Violations • Deterministic Enforcement • POSIX-Inspired**
+
+[![PyPI](https://img.shields.io/pypi/v/agent-os?label=PyPI)](https://pypi.org/project/agent-os/)
+[![License](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
+[![Python](https://img.shields.io/badge/python-3.10+-blue.svg)](https://python.org)
+
+[Quick Start](#-quick-start) • [Why Agent OS?](#-why-agent-os) • [Demos](#-live-demos) • [Integrations](#-integrations) • [Contributing](#-contributing)
+
+</div>
+
+---
+
+## 🎯 What is Agent OS?
+
+**Agent OS treats LLMs like raw compute and provides OS-level governance.** 
+
+Current AI agent frameworks let the LLM "decide" everything - including whether to follow safety rules. Agent OS inverts this: the **kernel decides**, the LLM just computes.
+
+```
+┌─────────────────────────────────────────────────────────┐
+│              USER SPACE (Untrusted LLM)                 │
+│   Your agent code runs here. It can crash, hallucinate, │
+│   or misbehave - the kernel survives.                   │
+├─────────────────────────────────────────────────────────┤
+│              KERNEL SPACE (Trusted)                     │
+│   Policy Engine │ Flight Recorder │ Signal Dispatch     │
+│   If agent violates policy → SIGKILL (non-catchable)   │
+└─────────────────────────────────────────────────────────┘
+```
+
+## ❓ Why Agent OS?
+
+### The Problem with Current Approaches
+
+| Approach | How it works | Why it fails |
+|----------|--------------|--------------|
+| **Prompt-based safety** | "Please don't do bad things" | LLM can ignore prompts |
+| **Constitutional AI** | Train model to be safe | Training doesn't guarantee |
+| **Guardrails** | Check output after generation | Too late - damage done |
+| **Agent OS** | **Kernel-level enforcement** | **Can't bypass the kernel** |
+
+### Real Benchmark Results
+
+| Metric | Prompt-based | Agent OS |
+|--------|-------------|----------|
+| Safety Violations | 26.67% | **0.00%** |
+| Response Tokens | 26.1 avg | **0.5 avg** |
+| Deterministic | No | **Yes** |
+
+## 🚀 Quick Start
+
+### Option 1: pip install (30 seconds)
+
+```bash
+pip install agent-os
+
+# Verify installation
+python -c "from agent_os import KernelSpace; print('Agent OS ready!')"
+```
+
+### Option 2: GitHub CLI Extension (Recommended for daily use)
+
+```bash
+# Install the gh extension
+gh extension install imran-siddique/gh-agent-os
+
+# Now use Agent OS directly from your terminal
+gh agent-os run "analyze this codebase for security issues"
+gh agent-os audit --policy pci-dss
+gh agent-os status
+```
+
+### Option 3: Docker (Production)
+
+```bash
+docker pull ghcr.io/imran-siddique/agent-os:latest
+docker run -it agent-os agentctl --help
+```
+
+## 💡 Core Concepts (5 minutes)
+
+### 1. Signals - Control Your Agents Like Processes
+
+```python
+from agent_os import AgentSignal, SignalDispatcher
+
+# Pause agent to inspect state (like Ctrl+Z)
+dispatcher.signal(agent_id, AgentSignal.SIGSTOP)
+
+# Resume execution
+dispatcher.signal(agent_id, AgentSignal.SIGCONT)
+
+# Policy violation? Kernel kills it (non-catchable)
+dispatcher.signal(agent_id, AgentSignal.SIGKILL)
+```
+
+### 2. Virtual File System - Memory That Makes Sense
+
+```python
+from agent_os import AgentVFS
+
+vfs = AgentVFS(agent_id="agent-001")
+
+# Standard mount points (like /home, /tmp, /etc)
+vfs.write("/mem/working/task.txt", "Current task...")  # Ephemeral
+vfs.write("/mem/episodic/session.log", "What happened")  # Experience
+vfs.read("/policy/allowed_actions.yaml")  # Read-only rules
+```
+
+### 3. IPC Pipes - Agents Talk Through Policies
+
+```python
+from agent_os.iatp import Pipeline, PolicyCheckPipe
+
+# Unix-style piping: agent1 | policy | agent2
+pipeline = Pipeline([
+    research_agent,
+    PolicyCheckPipe(allowed=["ResearchResult"]),  # Type check!
+    summary_agent
+])
+
+# If research_agent outputs wrong type → pipeline fails safely
+result = await pipeline.execute("Find recent AI papers")
+```
+
+## 🎬 Live Demos
+
+Four production-ready demos showing Agent OS in action:
+
+| Demo | Industry | What it shows | Run it |
+|------|----------|---------------|--------|
+| **[Carbon Auditor](examples/carbon-auditor/)** | Climate | CMVK drift detection catches $5M fraud | `python examples/carbon-auditor/demo.py` |
+| **[Grid Balancing](examples/grid-balancing/)** | Energy | 100 agents negotiate in 100ms | `python examples/grid-balancing/demo.py` |
+| **[DeFi Sentinel](examples/defi-sentinel/)** | Crypto | Stop hacks in 142ms | `python examples/defi-sentinel/demo.py` |
+| **[Pharma Compliance](examples/pharma-compliance/)** | Healthcare | Find contradictions in 100K pages | `python examples/pharma-compliance/demo.py` |
+
+```bash
+# Run all demos
+cd examples && docker-compose up
+```
+
+## 🔌 Integrations
+
+### Works With Your Existing Stack
+
+Agent OS doesn't replace your tools - it governs them:
+
+```python
+# With LangChain
+from agent_os.integrations import langchain_kernel
+chain = langchain_kernel.wrap(your_langchain_agent)
+
+# With CrewAI
+from agent_os.integrations import crewai_kernel
+crew = crewai_kernel.wrap(your_crew)
+
+# With AutoGen
+from agent_os.integrations import autogen_kernel
+autogen_kernel.govern(your_autogen_agents)
+```
+
+### GitHub Integration
+
+```yaml
+# .github/workflows/agent-os.yml
+name: Agent OS Safety Check
+on: [push, pull_request]
+
+jobs:
+  safety-audit:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: imran-siddique/agent-os-action@v1
+        with:
+          policy: .agent-os/policy.yaml
+          fail-on-violation: true
+```
+
+### VS Code Extension
+
+```bash
+# Install from marketplace
+code --install-extension imran-siddique.agent-os
+
+# Or use Command Palette: "Agent OS: Audit Current File"
+```
+
+### CLI for Daily Workflows
+
+```bash
+# Audit any agent codebase
+agentctl audit ./my-agent-project
+
+# Run with safety constraints
+agentctl run --policy strict "agent.py"
+
+# Monitor running agents
+agentctl status --watch
+
+# Replay agent execution from flight recorder
+agentctl replay --from checkpoint-001
+```
+
+## 📦 Package Architecture
+
+```
+agent-os/
+├── L1: Primitives (Foundation)
+│   ├── primitives     # Failure types, base models
+│   ├── cmvk           # Cross-model verification
+│   ├── caas           # Context-as-a-Service
+│   └── emk            # Episodic Memory Kernel
+│
+├── L2: Infrastructure (Communication)
+│   ├── iatp           # Inter-Agent Trust Protocol
+│   ├── amb            # Agent Message Bus
+│   └── atr            # Agent Tool Registry
+│
+├── L3: Framework (Governance)
+│   └── control-plane  # Kernel, signals, VFS
+│
+└── L4: Intelligence (Self-Correction)
+    ├── scak           # Self-Correcting Agent Kernel
+    └── mute-agent     # Reasoning/Execution split
+```
+
+Install what you need:
+
+```bash
+pip install agent-os              # Core only
+pip install agent-os[control-plane]  # + Governance
+pip install agent-os[full]        # Everything
+```
+
+## 🆚 How We Compare
+
+| Feature | LangChain | AutoGen | CrewAI | AIOS | **Agent OS** |
+|---------|-----------|---------|--------|------|--------------|
+| Multi-agent | ✅ | ✅ | ✅ | ✅ | ✅ |
+| Safety guarantees | ❌ | ❌ | ❌ | ❌ | **✅ Kernel-level** |
+| Deterministic | ❌ | ❌ | ❌ | ❌ | **✅** |
+| Process isolation | ❌ | ❌ | ❌ | ❌ | **✅ Kernel/User** |
+| Policy enforcement | ❌ | ❌ | ❌ | ❌ | **✅ SIGKILL** |
+| Audit trail | Partial | Partial | Partial | ❌ | **✅ Flight Recorder** |
+| Memory model | Ad-hoc | Ad-hoc | Ad-hoc | Short/Long | **✅ VFS** |
+
+**TL;DR**: Other frameworks help you *build* agents. Agent OS helps you *trust* them.
+
+## 🤝 Contributing
+
+We welcome contributions! Agent OS is designed to be extended.
+
+### Good First Issues
+
+- 🏷️ [`good-first-issue`](https://github.com/imran-siddique/agent-os/labels/good-first-issue) - Perfect for newcomers
+- 📚 [`documentation`](https://github.com/imran-siddique/agent-os/labels/documentation) - Help improve docs
+- 🧪 [`needs-tests`](https://github.com/imran-siddique/agent-os/labels/needs-tests) - Add test coverage
+
+### Development Setup
+
+```bash
+# Clone and setup
+git clone https://github.com/imran-siddique/agent-os.git
+cd agent-os
+pip install -e ".[dev]"
+
+# Run tests
+pytest
+
+# Run a specific demo
+python examples/carbon-auditor/demo.py
+```
+
+### Integration Bounties
+
+Want to add an integration? We're looking for:
+
+| Integration | Bounty | Status |
+|-------------|--------|--------|
+| LangChain adapter | 🎁 | Open |
+| CrewAI adapter | 🎁 | Open |
+| AutoGen adapter | 🎁 | Open |
+| OpenAI Swarm adapter | 🎁 | Open |
+| Anthropic Claude adapter | 🎁 | Open |
+
+See [CONTRIBUTING.md](CONTRIBUTING.md) for full guidelines.
+
+## 📖 Documentation
+
+- **[Architecture Guide](docs/architecture.md)** - Deep dive into kernel design
+- **[API Reference](docs/api/)** - Full API documentation
+- **[Examples](examples/)** - Working demos
+- **[AIOS Comparison](docs/AIOS_COMPARISON.md)** - Detailed competitor analysis
+
+## 🔬 Research
+
+Agent OS is designed for both production and academic use:
+
+- **Target**: ASPLOS 2026 Workshop on Agentic Systems
+- **Paper**: "Agent OS: A Safety-First Kernel for Autonomous AI Agents"
+- **Novel Contribution**: First OS-level governance for LLM agents
+
+## 📜 License
+
+MIT License - Use it, modify it, ship it. See [LICENSE](LICENSE).
+
+---
+
+<div align="center">
+
+**Built for engineers who don't trust their agents (yet).**
+
+[⭐ Star us on GitHub](https://github.com/imran-siddique/agent-os) • [📦 PyPI](https://pypi.org/project/agent-os/) • [💬 Discussions](https://github.com/imran-siddique/agent-os/discussions)
+
+</div>