HarperZ9/telos

Overview

HarperZ9/telos is an open-source Python framework for constructing shared AI workspaces that support creation, simulation, verification, and MCP tool integration. It records actions as replayable receipts, enabling deterministic traceability and collaborative debugging across AI agents.

Best for

Best for
Developers building collaborative AI agent systems that need deterministic replay and verification.

Use cases

• Build multi-agent workspaces with MCP tools for collaborative reasoning
• Run simulations with reproducible receipts to verify AI agent behavior
• Create shared environments for debugging and replaying agent interactions

How to use

Install

node demo/run.mjs

Tools exposed

• telos.*
• gather
• index
• forum
• crucible
• ci-doctor.mjs
• ci-triage.mjs
• presentation-doctor.mjs
• accessibility-doctor.mjs
• performance-doctor.mjs
• compatibility-doctor.mjs
• operator-doctor.mjs
• mcp-freshness.mjs
• proof.mjs
• showcase.mjs
• context-envelope.mjs
• context-pack.mjs
• action-receipt.mjs
• loop-ledger.mjs
• creative-engine.mjs

Tested with

Codex, Claude, OpenAI Agents

Example client config

{\n  "servers": {\n    "telos-mcp": {\n      "transport": "stdio",\n      "host": "localhost",\n      "port": 10000\n    }\n  }\n}

Notes

HarperZ9/telos is an open-source Python framework for constructing shared AI workspaces that support creation, simulation, verification, and MCP tool integration. It records actions as replayable receipts, enabling deterministic traceability and collaborative debugging across AI agents.

3 stars on GitHub. Last updated 2026-07-07.

Use cases

• Build multi-agent workspaces with MCP tools for collaborative reasoning
• Run simulations with reproducible receipts to verify AI agent behavior
• Create shared environments for debugging and replaying agent interactions

Pros

• Open-source and Python-based, easy to integrate into existing agent stacks
• Replayable receipts provide concrete audit trails for agent actions
• Supports MCP tools for extensible agent ecosystems

Cons

• Small community (3 stars), limited third-party plugins or documentation
• Primarily focused on Python, not language-agnostic for polyglot teams
• Narrowly scoped to shared workspaces, not a general-purpose agent framework

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.