Agent♥︎Age
Catalog

io.github.codeledgerECF/codeledger

Official

by codeledgerECF · Shell

Local context and engineering evidence for AI coding agents.

io.github.codeledgerECF/codeledger Model Context Protocol (MCP) Server

This MCP server provides deterministic context selection for AI coding agents. It maintains local memory that compounds across sessions, scoring files in a repository and selecting only what the current task requires, then capturing outcomes and promoting successful patterns.

🛠️ Key Features

  • Deterministic context control plane
  • Scores files in a repository
  • Selects context based on current task needs
  • Captures outcomes
  • Promotes successful patterns into reusable institutional memory
  • Local memory that compounds across sessions

🚀 Use Cases

  • Reducing irrelevant files consumed by an AI coding agent’s context window
  • Managing “cold start” sessions for agents by retaining institutional knowledge
  • Pre-merge risk signaling via controlled context selection (as described)

⚡ Developer Benefits

  • More efficient use of the context window by selecting task-relevant files
  • Reuse of successful development patterns across sessions

⚠️ Limitations

  • The provided description does not specify supported tools, integrations beyond “github-app,” or configuration details.

Topics

codeledgergithub-appcontext-checkagent-governanceai-coding-agentclaude-codeclicode-reviewcodexcoding-agentscontext-engineeringcontext-managementcursordeveloper-toolsdevexgemini-cligithub-copilotlocal-firstmcpsoftware-engineering