Show HN: Selvedge – an MCP server that captures why AI agents change code (selvedge.sh)

Selvedge has introduced an innovative MCP server that enhances AI coding by capturing the rationale behind modifications made by AI agents in real-time. When AI tools like Claude Code, Cursor, or Copilot make changes, Selvedge records the contextual reasoning for each alteration, storing this information in a local SQLite database. This feature ensures that developers can understand "why" a change was made, even long after the AI session has ended, addressing the common issue where context is lost over time. The significance of Selvedge lies in its ability to provide long-term memory for AI-coded projects, allowing teams and solo developers to query not just what changes occurred, but why they were implemented. By maintaining a clear record of the decision-making process behind every code update, Selvedge functions as a crucial layer of provenance that complements existing version control and observability tools without replacing them. This capability is particularly valuable for projects with long lifespans, where developers need to revisit AI contributions, ensuring clarity and accountability in both collaborative and individual coding efforts. As an open-source project, Selvedge aims to enhance AI-assisted development practices, making it easier to maintain and evolve codebases over time.