ECC
A coding assistant toolkit for Claude Code and Codex, with local Markdown memory, MIT-licensed workflows, and support for self-hosted model endpoints.
ECC is a free, MIT-licensed open source toolkit that adds repeatable engineering workflows to coding assistants. It runs alongside your chosen agent and works with self-hosted models or cloud providers through that agent's supported endpoints. It's for developers who want planning, testing and review to follow a consistent process across coding sessions.
Claude Code has the fullest support. ECC also supports Codex, while adapters for Cursor, OpenCode, Gemini, Zed and other editors have fewer capabilities. ECC supplies the workflows; the agent and its model backend determine where inference runs. The separate ECC Pro product is a hosted GitHub App.
Its workflows produce editable plans before implementation, guide test-driven development through checks backed by evidence, and use a separate review context to look for regressions. Specialized agents cover build repair, security and architecture work. Skills load when a task needs them, while hooks can enforce checks outside the model's prompt. AgentShield scans agent configuration, including prompts, permissions, secrets and MCP settings.
The Memory Vault stores durable context in local, inspectable Markdown files that different agents can share. It keeps project, team and personal memories separate, with session summaries and learned patterns rather than copied conversation transcripts. Memories remain unreviewed context rather than executable rules. ECC doesn't upload diagnostics automatically, and its desktop dashboard lets you search and browse agents, skills, commands and rules.
