TRACE
Active·★ 25·Apache-2.0·Updated 2026-10-13
TRACE is a protocol for auditing decision provenance in AI-assisted scientific workflows. It records who proposed what, human or AI, and does not allow the AI to resolve its own proposal.
#ai-transparency#audit-trail#claude-code#human-ai-collaboration#mcp#mcp-servers#model-context-protocol#open-science
01
Features
01Event log of proposals, resolutions and retractions with provenance
02AI is not permitted to resolve its own proposals
03Works with Claude through MCP in a scoped session
02
Compatibility
Claude Code
Client
Verified via docs
Cursor
IDE / Editor
Verified via docs
Codex
Client
Verified via docs
Python
Runtime
Verified via docs
Cross-platform
OS
Verified via docs
03
Quick start
1
$ uv pip install -e ".[dev]"
04
Use cases
↳Audit who made each methodological decision in a research project
↳Keep a transparent record of human and AI contributions
05
Alternatives
agent-bridge★ 415
A local bridge for bidirectional collaboration between Claude Code and Codex. 连接 Claude Code 与 Codex 的本地实时协作桥接工具。
obsidian-mcp-connector★ 119
Add integrations like semantic search and custom Templater prompts to Claude or any MCP client.
ruvos★ 28
rUvOS — the agentic operating system. A single static Rust binary that runs as an MCP server, giving Claude Code persistent memory, resumable signed sessions, multi-agent orchestration, a knowledge graph, safety guardrails, and a signed audit trail. Zero Node.js, zero SQLite, zero daemon.
call-me★ 25
Let your AI agents call and text your actual iPhone — Claude Code plugin + agent skill
snipeit-mcp★ 36
MCP server for Snipe-IT asset management. Enables AI assistants to manage assets, consumables, accessories, licenses, users, and system configuration via 39 comprehensive tools.
Comments
Log in to leave a comment
No comments yet. Be the first!