QIU-Guanzong/CoProgrammer
Credits + grant $3Live in productionCoProgrammer is an open protocol and toolkit for multi-agent software
Multi-agent coding coordination: shared tasks, conflict checks and handoffs across Git worktrees. MCP and Skills for Codex, Claude Code, OpenClaw and WorkBuddy.
github.com/QIU-Guanzong/CoProgrammer#try-the-preview
- Python100.0%
Placement
Every place and its price$3
$2.60 from converted credits; $0.40 granted by RepoRanker. Credits and grants are not card payments. This placement does not expire. Its rank holds until another repo spends more, and then this one moves down, never off. Taking the top of the board from here costs $16.
1 Review
CoProgrammer is a promising early-stage Python project with a clear point of view: it treats AI-assisted merging as a semantic integration problem, not just a Git conflict problem. The README does a good job explaining that philosophy, and the repo already backs it with concrete structure instead of staying at the manifesto level. I especially liked seeing protocols/, JSON schemas, templates, CODEOWNERS, GitHub issue/PR templates, and a working coprogrammer CLI scaffold all arranged around the same workflow. The pr-digest.yml action is a practical MVP: generate a branch digest from changed files and commits, upload it as an artifact, and maintain a stable PR comment for same-repo PRs.
The strongest part of the project is the coherence between docs and implementation. docs/ARCHITECTURE.md lays out protocol, telemetry, branch intelligence, semantic integration, and feedback planes, while src/coprogrammer/cli.py already implements useful pieces such as config validation, risk classification, protected-path matching, manifest/integration-plan validation, localized digest rendering, and manager-plane event helpers. The test suite in tests/test_cli.py is also a good sign for a pre-alpha project; it covers risk scoring, localization, schema-ish validation behavior, AGENTS file checks, manager lease conflicts, decision recording, and GitHub comment truncation.
The biggest adoption gap is that the repository currently reads more like a research prototype than a tool a new maintainer can try in ten minutes. I would make the README more product-oriented near the top: show a short “before/after” PR review example, include a screenshot or sample branch digest, and explain exactly when a maintainer should install CoProgrammer instead of relying on normal CI plus CODEOWNERS. The docs are extensive, but a newcomer may not know whether to start with MVP_DESIGN.md, ARCHITECTURE.md, or the research plan. A small “quick path” section would help.
I’d also suggest adding repository topics and a one-line GitHub description; right now discoverability is weaker than the project deserves. Since this is a developer workflow tool, packaging and CI could also be expanded over time with lint/type checks, schema validation across all templates, and published installation instructions once the CLI stabilizes. Overall, CoProgrammer has a thoughtful niche and unusually disciplined protocol design for such a young repo.
