quint-code

Roadmap

quint-code is actively developed. Here's what's coming — not promises, but directions we're building toward.

Recently shipped (5.3)

  • Interactive terminal dashboard — quint-code board with four tabs, vim navigation, glamour markdown, live refresh, health checks
  • Context-aware /q-reason — three paths based on intent (think, prepare-and-wait, or full autonomous)
  • Decision mode from artifact chain — mode reflects actual reasoning depth, not agent's self-report
  • C/C++ language support — compile_commands.json, header modules, symlink-safe resolution
  • FTS5 search enrichment — agent-generated keywords at write time

Team collaboration

This is the #1 priority. The architecture is already prepared:

  • Database is out of the repo (unified storage in ~/.quint-code/)
  • Project identity via .quint/project.yaml — stable across machines
  • QUINT_SERVER_ORIGIN env var ready for remote server mode
  • Markdown projections in .quint/ — git-mergeable, reviewable in PRs
  • quint-code board — the foundation for team-visible dashboards

The plan for team mode:

quint-code server — a new subcommand in the same binary. A thin HTTP server that manages the decision database. For solo advanced use, it runs on localhost and enables features that require a long-running process. For teams, it runs on a shared server. 

Solo:  Agent → MCP → quint-code serve → local SQLite
Team:  Agent → MCP → quint-code serve → quint-code server (remote) → PostgreSQL

When team mode exists, quint-code board switches to the remote database automatically. quint-code board --check becomes a real CI command that connects to the server.

Formal specification (research)

Decision records from the full cycle contain invariants, post-conditions, and acceptance criteria — enough structure for any agent to generate property-based tests today. This works without any quint-code changes.

The research direction: can we go further? Translating decision invariants into Quint specifications (TLA+ with TypeScript syntax) for model checking — verifying that invariants hold across all reachable states, not just random inputs. This is in research phase.

Cognitive review loops

Instead of dumb valid_until timers, use FSRS (Free Spaced Repetition Scheduler) to schedule decision reviews based on evidence confidence and historical review quality. Requires the server mode (long-running process).

Autonomy budgets

From FPF E.16: treat agent autonomy as a managed resource (like money or risk), not as "magic of a smart model." Typed budgets with guards, separation of duties, override protocols. This is the team governance layer.

What's not on the roadmap

  • Cloud/SaaS — quint-code is local-first. If there's ever a hosted version, it'll be a server you run yourself, not a SaaS.
  • AI model lock-in — quint-code works with any MCP-compatible AI tool. We don't pick your model.
  • Prescriptive methodology — quint-code provides structure, not opinions. Use as much or as little of the FPF cycle as fits your workflow.

Want to contribute?

Check the issues on GitHub. The contributing guide explains the workflow: issue first, then PR to dev.