telos-framework

Telos: Purpose-Driven Multi-Agent Development Framework

A philosophically-grounded AI collective that embeds Aristotelian purpose hierarchy into software development. Every line of code serves your ultimate mission.

Philosophy

Telos (Greek: τέλος) means "end," "purpose," or "goal"—the ultimate reason something exists.

Logos (Greek: λόγος) means "reason," "discourse," or "rational principle"—the organizing intelligence that maintains coherent order.

Once you install the Telos framework, your AI coding assistant becomes the Logos orchestrator itself. Every conversation with your AI follows bidirectional validation: strategic purpose flows downward (L9→L1) to guide implementation, while technical constraints flow upward (L1→L9) to inform strategy. Your AI assistant mediates this dialogue, ensuring decisions converge only when both flows align.

Most development tools focus on how to code. Telos ensures you're building the right thing by maintaining alignment between implementation and purpose across nine levels of abstraction—from syntax to transcendent meaning.

Based on Kenneth Boulding's hierarchy of system complexity, Telos orchestrates specialized AI agents that operate coherently from low-level linting to strategic vision.

Architecture

The Nine Levels

L9: Telos-Guardian      → Strategic alignment with ultimate purpose
L8: Market-Analyst      → Business metrics and KPIs
L7: Insight-Synthesizer → User feedback and behavioral analytics
L6: UX-Simulator        → User experience and accessibility
L5: Journey-Validator   → End-to-end workflows and integration
L4: Integration-Contractor → API contracts and service boundaries
L3: Component-Architect → Component design and composition
L2: Function-Author     → Unit logic and TDD
L1: Syntax-Linter       → Code structure and formatting

Logos Orchestrator

The central orchestration engine implementing:

• Top-down decomposition: Strategic goals cascade into tactical specs
• Bottom-up validation: Implementation validates against purpose
• Middle-out reconciliation: Conflicts resolved through rational dialogue
• Spec-driven workflow: All changes flow through OpenSpec proposals

Quick Start

npx telos-framework init

This command:

• Detects your AI coding platform (Claude Code, Opencode, Cursor, etc.)
• Installs platform-specific slash commands
• Sets up configuration files (CLAUDE.md, AGENTS.md, etc.)
• Prepares your project for Logos-driven development

Note: Safe to run multiple times! If Telos is already installed, you'll be prompted to choose: Abort (keep existing), Reinstall (overwrite), or Add platforms. See Re-initialization Safety for details.

After installation, open your project in your AI coding assistant and run /telos-init to analyze your codebase and generate the 9-level purpose hierarchy.

Features

Adaptive Tool Integration

Telos discovers and integrates your existing tools:

• Linters (ESLint, Ruff, etc.)
• Test frameworks (Vitest, Jest, Playwright)
• Analytics (PostHog, Amplitude)
• MCP servers
• Custom tooling

Multi-Platform Support

Works with:

• Claude (Code, Projects)
• Cursor
• GitHub Copilot
• Google Gemini
• Any AI coding assistant

Single source of truth with platform-specific symlinks—no duplication.

OpenSpec Integration

Full integration with OpenSpec workflow:

• Logos creates proposals for changes
• Agents work through specs and tasks
• Validation uses OpenSpec archive
• Telos lineage tracked throughout

Technical Agent Delegation

9-level agents can delegate to specialized technical agents from Claude Collective:

• research-agent for technical investigation
• quality-agent for comprehensive reviews
• testing-implementation-agent for test suites
• devops-agent for deployment
• And more specialized agents

Commands

CLI Commands

telos init              # Install Telos slash commands and memory files
telos --help            # Show all commands

Slash Commands (AI-Native)

Use these in Claude Code, OpenCode, or Cursor:

/telos-init             # Initialize Telos with AI-driven analysis
/telos-quick            # Fast initialization (auto-accept AI proposals)
/telos-validate         # Check code alignment with purpose hierarchy
/telos-status           # Show current Telos configuration
/telos-reset            # Clear and reinitialize

Project Structure

After initialization:

your-project/
├── telos/
│   ├── content/
│   │   ├── TELOS.md         # Your project's purpose hierarchy
│   │   ├── AGENTS.md        # Consolidated agent definitions
│   │   ├── LOGOS.md         # Orchestrator instructions
│   │   └── TOOLS.md         # Tool registry and mappings
│   ├── agents/              # Individual agent definitions (L1-L9)
│   └── templates/           # Platform-specific configs
├── .telos/                  # Runtime state (gitignored)
├── TELOS.md                 # Symlink for visibility
└── [platform symlinks]      # CLAUDE.md, .cursor/rules/, etc.

Philosophy & Theory

Why Hierarchical Agents?

Flat agent collectives lack governance—agents can work at cross-purposes. Hierarchical organization with clear level boundaries ensures:

• Strategic coherence (top-down)
• Implementation integrity (bottom-up)
• Efficient specialization (right tool, right level)

Why Telos?

Without explicit purpose capture, AI assistants optimize for immediate requests, not ultimate goals. Telos makes purpose explicit and traceable, enabling:

• Alignment validation at every level
• Conflict resolution through purpose appeal
• Emergent strategic consistency

Why Logos?

Unstructured agent communication devolves into noise. Logos enforces rational discourse through:

• Spec-driven dialogue (OpenSpec format)
• Structured reporting protocols
• Level-appropriate context filtering

Examples

See /examples for:

• Simple web app initialization
• Existing codebase integration
• Multi-platform usage demonstration

Contributing

See CONTRIBUTING.md for how to contribute to this project.

Philosophy Deep Dive

See PHILOSOPHY.md for comprehensive explanation of:

• Aristotelian teleology in software
• Boulding's hierarchy applied to development
• Logos as rational agent orchestration
• Ontological levels and emergence

Troubleshooting

See TROUBLESHOOTING.md for common issues and solutions.

License

MIT License - see LICENSE

Transform AI-assisted development from reactive tool usage to coherent, purpose-aligned creation.

GitHub Repository | Documentation | Report Issues