convez.ai
MCP ServerAutomaçãoAvançado
DDD-Enterprises

Dopemux MVP

DDD-Enterprises/dopemux-mvp

Espaço de trabalho operador ancorado em repositório com múltiplas autoridades

Ver no GitHub
BomPluginScore · de 0 a 100

Stack de operações cognitivas consciente de ADHD para desenvolvedores, que funciona como uma superfície de controle orientada ao operador para fluxos de trabalho com autoridade dividida. Coordena trabalho através de múltiplos sistemas mantendo a fonte de verdade visível para cada ação, sem consolidar PM, memória, recuperação ou execução em um único proprietário.

Instalação

$npx convez-ai DDD-Enterprises/dopemux-mvp

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
3
Forks
0
Acompanhando
0
Saúde
87%
Último commit
há 1 dia
Versão
—
Issues abertas
90
Contribuidores
6

Casos de uso

  • Coordenar desenvolvimento em múltiplos sistemas mantendo limites de autoridade claros
  • Gerenciar workflows com transições explícitas entre diferentes domínios (PM, execução, memória, auditoria)
  • Suportar times neurodivergentes com operações cognitivas estruturadas e visibilidade de contexto
  • Implementar padrões de split-authority em repositórios complexos

Veredito

Dopemux é uma arquitetura sofisticada para operações em repositórios com autoridades distribuídas, bem documentada e rigorosamente boundary-aware. Ideal para times que precisam de coordenação explícita entre sistemas heterogêneos sem pretender unificar verdades ortogonais.

README

A tradução para o português deste README ainda está na fila.

Dopemux MVP

Dopemux is an operator-facing control surface for split-authority development workflows.

It exists for work that crosses local startup, execution handoff, PM workflow, structured context, chronicle history, derived retrieval, bridge routing, operator support, and repo audit without pretending one system owns every truth lane.

What Dopemux Is

Dopemux is a repo-grounded operator workspace. It helps an operator coordinate development work across multiple systems while keeping the source of truth visible for each action.

What Dopemux Is Not

Dopemux is not one AI assistant, one PM platform, one memory system, or one autonomous development brain. It coordinates work across authority lanes. It does not collapse those lanes into one owner.

Why It Exists

The repository contains real operational systems, but the authority model is split. PM metadata, workflow transitions, decisions, progress, chronicle receipts, retrieval, bridge transport, operator support, and repo-truth audit have different boundaries. Dopemux keeps those boundaries visible so operators can move work without turning derived views, mirrors, or proxy routes into source truth.

This README is repo-grounded as of 2026-05-19. Runtime code, config, compose wiring, tests, and active entrypoints remain stronger than this document. Where the repository does not prove a claim, the claim is marked UNKNOWN or NEEDS_REPO_VERIFICATION.

Where Humans Should Start

Where AI Tools Should Start

Authority Model

Treat authority by domain, not by brand name:

DomainStrongest observed authorityDo not overclaim
Operator control and startupdopemux CLI in src/dopemux/cli.py and command modulesDoes not own PM, durable memory, retrieval, or execution after handoff.
Execution after handoffExternal dopetask via scripts/dopetaskscripts/taskx is only a compatibility shim.
PM metadataLeantime through dopemux PM adaptersWorkflow legality is not Leantime's slice.
Workflow transitionstask-orchestrator workflow surfacestask-orchestrator does not own all PM state.
Decisions, progress, structured contextConPortConPort is not all memory and not PM metadata authority.
Historical receipts and chronicledope-memory chronicle ledgerdope-memory is not current PM state authority.
Code/docs retrievaldope-context and ConPort retrieval surfacesRetrieval output is derived, not source truth.
Bridge, proxy, event transportdopecon-bridgeBridge routes are not source truth.
Operator supportADHD EngineADHD Engine is not PM or memory authority.
Repo audit/extractionRepo Truth Extractor runtimeExtraction artifacts do not outrank runtime.
Agent ownershipUNKNOWNDo not document a single repo-wide agent authority without a runtime pass.

Quick Start

For the short operator path, see QUICK_START.md. The documentation tutorial version is docs/01-tutorials/quickstart.md.

Minimal setup:

git clone https://github.com/DDD-Enterprises/dopemux-mvp
cd dopemux-mvp
uv sync --frozen --extra dev

Manual compose startup uses the canonical root compose file:

docker network inspect dopemux-network >/dev/null 2>&1 \
  || docker network create dopemux-network
docker compose -f compose.yml up -d --build

Then run the operator CLI:

dopemux start

dopemux start is the operator cockpit entrypoint. The inspected CLI code shows routing, MCP/server coordination, agent validation, context, and launch behavior. It does not prove that the full compose stack is created for every profile; use the explicit compose command above when you need the compose services.

Default Local Service Checks

The observed compose and registry defaults include:

ServiceDefault host portHealth path
dopecon-bridge3016/health
ConPort HTTP3004/health
dope-context3010/health
dope-memory3020/health
task-orchestrator8000/health
ADHD Engine3025/health
Leantime8080/
Qdrant6333/

Example checks after compose startup:

curl -fsS http://localhost:3016/health
curl -fsS http://localhost:3004/health
curl -fsS http://localhost:3010/health
curl -fsS http://localhost:3020/health
curl -fsS http://localhost:8000/health
curl -fsS http://localhost:3025/health

These are compose-backed defaults, not proof that every local profile uses the same overrides. If a local .env changes ports, follow the runtime environment and record the override.

Common Workflows

  • Start the operator cockpit: dopemux start
  • Inspect CLI command groups: dopemux --help
  • Delegate kernel execution: dopemux kernel <command>
  • Use the canonical Repo Truth Extractor command family: dopemux rte ...
  • Use legacy RTE compatibility only when needed: dopemux upgrades ...
  • Start compose services manually: docker compose -f compose.yml up -d --build
  • Check compose service state: docker compose -f compose.yml ps
  • Validate docs before closing documentation work:
bash scripts/lint-docs.sh
python scripts/docs_validator.py
python scripts/docs_frontmatter_guard.py
python scripts/check_root_hygiene.py
git diff --check

Documentation Map

Known Drift And Limits

  • task-orchestrator runtime authority is still conflicted across active app code, Docker/startup wiring, and older references. Do not declare that drift closed without runtime validation.
  • ConPort surfaces are split across HTTP/API and MCP/SSE style ports. Use the port and route that matches the operation.
  • dope-memory and working-memory-assistant naming overlap. Treat dope-memory as chronicle authority unless a specific runtime path proves more.
  • dopecon-bridge exposes broad PM and knowledge-graph-looking routes, but it is still bridge/proxy/event transport only.
  • dope-context retrieval output is useful context, not source truth.
  • Repo Truth Extractor artifacts are evidence outputs, not stronger than the runtime they describe.
  • Broader agent runtime ownership is UNKNOWN.
  • Some live startup and health behavior remains NEEDS_REPO_VERIFICATION because this documentation packet does not start the full runtime stack.

Contributor Notes

  • Read AGENTS.md before non-trivial repo work.
  • Use an active Task Packet for scoped repo-changing work.
  • Preserve the packet allowlist and avoid runtime code, service code, compose, dependency, or test edits unless a packet explicitly allows them.
  • Inspect runtime/source truth before editing docs.
  • Preserve contradictions and UNKNOWN instead of smoothing drift into a cleaner story.
  • Run narrow validators first, then broader validators when the packet requires them.

Status And License

The package metadata currently marks this project as alpha and MIT-licensed in pyproject.toml. The repository contains operational code and active docs, but this README does not claim production readiness.

Ver o original no GitHub

Bom para

Relacionados