What are 'Reasoning Traces' in Scribe?

Every log entry must include a 'why', 'what', and 'how' block to explain the decision-making process, active constraints, and methodology used by the agent.

Which documents are managed by Scribe?

Scribe manages the core project suite including the ARCHITECTURE_GUIDE.md, PHASE_PLAN.md, and CHECKLIST.md files via the manage_docs tool.

What does the Scribe MCP skill do?

Scribe acts as a documentation guardrail for AI agents, ensuring they log implementation details every three edits and follow structured planning protocols.

How does Scribe prevent technical debt?

It enforces strict rules against creating parallel or replacement files, requiring agents to refactor and improve existing code directly within the current infrastructure.

Does Scribe support semantic search?

Yes, Scribe uses vector-based semantic retrieval to allow agents to find relevant entries in documentation and progress logs based on intent.

Scribe Documentation Governance

Name: Scribe Documentation Governance
Author: CortaLabs

byCortaLabs

0•

协作工具

Enforces documentation governance and auditability for AI agents by managing project structure, progress logging, and reasoning traces.

Scribe is a documentation governance system designed to keep AI agents aligned, accountable, and auditable while building software. It serves as a set of guardrails that prevents agents from bypassing existing infrastructure or losing implementation context. By requiring agents to log every significant decision using a 'Three-Part Framework' (Why, What, How), Scribe creates a persistent chain of reasoning. It automates the creation and maintenance of architecture guides, phase plans, and checklists, ensuring that all AI-driven development is traceable and follows established best practices.

主要功能

01Structured management of Architecture Guides, Phase Plans, and Checklists

02Semantic search for retrieving project history and documentation context

03Strict guardrails against technical debt and parallel file creation

04Automated progress logging with mandatory reasoning traces (Why/What/How)

05Readiness diagnostics via Scribe Doctor to validate repo and plugin status

060 GitHub stars

使用场景

01Maintaining an auditable trail of AI decision-making for compliance and debugging

02Bootstrapping new software projects with standardized planning and architecture docs

03Enforcing strict project management protocols across multi-agent AI workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cortalabs/scribe_mcp scribe-mcp-usage

For use in Claude.ai and ChatGPT

Download Skill