Meronq reports completion of Sprints 1 through 4, building a shared engineering-data workflow that moves from repository inspection to structured “CEM” representations and local memory. Sprint 1 focuses on defining the Canonical Engineering Model (CEM) primitives used across components, including entity, relation, evidence, event, action, and result models in @meronq/core. It also updates the handshake so AI clients receive CEM snapshots and sprint hints at session start, with provisions for workspace discovery and a requirement to gate Git write access.
Sprint 2 then delivers a local project intelligence prototype. It introduces a scanner that builds a ProjectIndex from repository layout, converts it into a CEM snapshot using project data plus git log, and saves results to a local JSON file. CLI commands (such as meronq scan and meronq info) support this flow, and the local MCP handshake is updated to include project_scan and project_index.
Sprint 3 (described as part of the progression) adds a local “memory engine,” moving persistence toward SQLite.
Sprint 4 closes with a GitHub integration package (@meronq/github), adding GitHub repository metadata, issue and pull request data via the GitHub REST API, CEM translation, local MCP v1.13.0 modules, a local JSON cache, and CLI commands for syncing, status, summarization, and doc drafts.