Roadmap

English · Русский

These are directions, not committed dates. Priorities follow user feedback and confirmed defects. Stable 1.6.0 is already published and publicly verified; completed work belongs in release history, not this list.

Now

  • Process external stable 1.6.0 feedback through the OpenSpec work plan: reviewed numeric CSV profiling, authenticated Trino access, relational planning and utility evidence, bounded statistics and actionable CLI diagnostics. Preserve conservative privacy defaults and read-only database budgets. Start with reproduced failures; do not reopen corrected historical harness findings.

  • Improve onboarding and task-oriented documentation, including versioned and downloadable guides.

  • Run product-validation pilots on development and analytics workflows.
  • Address confirmed client defects, migration feedback and dependency maintenance.
  • Monitor source/output consistency, portable business rules and MCP permissions.

Next

  • Design batch-based validation of CSV, JSON and Parquet so complete datasets need not remain in memory. Bound cross-table key and uniqueness checks, preserve deterministic results and measure memory on synthetic fixtures. A separate OpenSpec change is required before implementation.
  • Simplify CSV and database workflows using pilot feedback.
  • Publish runnable examples and short case studies using synthetic fixtures.
  • Improve diagnostics where users repeatedly need help.

Later

  • Add integrations only after demonstrated demand.
  • Consider reporting and validation adapters for concrete workflows.
  • Evolve public contracts through the compatibility and deprecation policy.

The finite fictional transformation/session campaign for stable 1.6.0 completed with explicit coverage limits. That grants no dataset permission. Changed public activation still requires matching specifications, executable evidence and independent exact-source review.