Skip to content

GameDesignOS v1.2.0 — Intent Work Order & Workflow Governance

Latest
Latest

Choose a tag to compare

View all tags
@DY-2026 DY-2026 released this 23 Jul 08:31
v1.2.0
7eeb6f5

Highlights

v1.2.0 upgrades GameDesignOS from RJR-AI authority coverage into an intent-driven workflow governance system.

The core shift:

AI work starts from intent, not instructions.
Workflow runs preserve governance refs, not only node state.
paranoia-ai-system-evolver participates as a cross-workflow checkpoint, not a mega-agent.
Domain skills keep owning their specialist outputs.

Added

  • Added intent-work-order.schema.json for upgrading AI work orders from instruction sheets into intent work orders.
  • Added Intent Work Order playbooks and templates to paranoia-ai-system-evolver.
  • Added workflow-run.governance to the workflow-run contract.
  • Added Project Workflow Governance playbooks and workflow_governance_review templates.
  • Added Paranoia Checkpoints to Decision-to-Information, Idea-to-Validation, Media-to-Diagnosis, Weekly ED Experiment, and Evidence-to-Proposal workflows.
  • Added router coverage for overall process, project process, process improvement, workflow governance, output quality, and workflow-governance review requests.
  • Added behavior eval coverage for skipped workflow governance.

Changed

  • Runtime/package version is now 1.2.0.
  • Project-Ready workspace schema remains 1.0.0; no v1 workspace migration is required.
  • New workflow runs include a governance block in shadow mode by default.
  • README trio, runtime docs, package metadata, v1 workspace manifest, project-workspace contract, tests, and validators are synced to v1.2.
  • paranoia-ai-system-evolver now explicitly routes to the Intent Work Order and Project Workflow Governance playbooks.

Safety

  • Governance defaults to shadow; it records intent, VOI/RJR, drift, Human Gate, rollback, and candidate learning without blocking normal domain workflows.
  • paranoia-ai-system-evolver does not replace game-concept-architect, game-experience-analyzer, game-experience-density-optimizer, or game-design-proposal-writer.
  • Candidate learning still requires eval evidence, Human Gate, and rollback before promotion to durable rules.
  • v1.2.0 is tagged at the reviewed release commit. Follow-up schema-hardening and CLI fixes remain post-release changes until a later release is cut.

Validation

Recommended checks:

python scripts/validate_repo.py
python scripts/validate_skill.py paranoia-ai-system-evolver
python paranoia-ai-system-evolver/quick_validate.py paranoia-ai-system-evolver
python -m unittest discover -s scripts/tests
python -m gamedesignos --version
python -m gamedesignos ask "paranoia-ai-system-evolver 需要更多的参与到这个项目的流程中，优化整体流程和产出"

Tagging

The historical v1.0.0 and v1.1.0 tags and the v1.2.0 release tag now exist. Changes after v1.2.0 belong under Unreleased until the next reviewed release.

Rollback

If v1.2 must roll back:

  1. Keep v1 workspaces in place; schema_version did not change.
  2. Revert pyproject.toml, gamedesignos.constants.RUNTIME_VERSION, and v1 manifest runtime declarations to 1.1.0.
  3. Keep Intent Work Order and workflow-governance docs as candidate notes if they need more real-project validation.
  4. Remove the workflow-run.governance requirement from new workflow runs only if the governance block causes verified downstream breakage.
Assets 2
Loading