GitHub Release Checklist

Use this checklist before publishing publicly or cutting a major release. For the step-by-step maintainer flow, use docs/RELEASING.md.

1) Repository Basics

• [ ] README.md is current and links key docs.
• [ ] CONTRIBUTING.md exists and reflects current workflow.
• [ ] CODE_OF_CONDUCT.md exists.
• [ ] SECURITY.md exists with vulnerability reporting path.
• [ ] License is set for repository publication.
• [ ] CITATION.cff exists for academic referencing.

2) Documentation Completeness

• [ ] docs/DOCS_INDEX.md is up to date.
• [ ] Language spec and semantics are current (docs/AINL_SPEC.md, SEMANTICS.md).
• [ ] Training and eval docs match scripts:
  • docs/FINETUNE_GUIDE.md
  • docs/TRAINING_ALIGNMENT_RUNBOOK.md
• [ ] AI continuity/handoff docs are current:
  • docs/AI_AGENT_CONTINUITY.md
  • docs/CONTRIBUTING_AI_AGENTS.md
• [ ] Changelog includes this release (docs/CHANGELOG.md).

3) Quality and Reproducibility

• [ ] Core tests pass (.venv/bin/python scripts/run_test_profiles.py --profile core).
• [ ] Training/eval scripts parse and run in expected environment.
• [ ] Latest run artifacts are present and reviewable:
  • corpus/curated/model_eval_report_v5_aligned.json
  • corpus/curated/model_eval_trends.json
  • corpus/curated/alignment_run_health.json
• [ ] Primary quality metrics are reported:
  • strict_ainl_rate
  • runtime_compile_rate
  • nonempty_rate

4) GitHub Community UX

• [ ] Issue templates exist for bug reports and feature requests.
• [ ] PR template exists with test/docs checklist.
• [ ] CI workflow is green and documented.

5) AI-Led Attribution Clarity

• [ ] README clearly states human-initiated + AI-led co-development model.
• [ ] docs/PROJECT_ORIGIN_AND_ATTRIBUTION.md is linked and current.
• [ ] Attribution language is consistent across top-level docs.
• [ ] tooling/project_provenance.json matches current initiator/provenance metadata.
• [ ] docs/PROVENANCE_AND_RELEASE_EVIDENCE.md has been reviewed for this release.
• [ ] Release post(s) include the same initiator references and release commit/hash.

6) Coordination and Advanced Surfaces

• [ ] Coordination protocol lock tests pass (e.g. tests/test_agent_protocol_surface.py).
• [ ] Coordination baseline artifacts and docs are in sync (see docs/advanced/AGENT_COORDINATION_CONTRACT.md baseline section).
• [ ] Coordination mailbox validator has been run on baseline task/result artifacts (e.g. python -m scripts.validate_coordination_mailbox --tasks-jsonl ...).
• [ ] README.md and docs/DOCS_INDEX.md still clearly separate core/safe-default surfaces from advanced/operator-only/experimental surfaces.
• [ ] Advanced coordination/OpenClaw examples remain labeled as extension-only, advanced, and advisory-only in docs/EXAMPLE_SUPPORT_MATRIX.md and their file headers.

Advanced memory and TTL hygiene:

• [ ] Memory adapter contract and verbs (put, get, append, list, delete, prune) match docs/adapters/MEMORY_CONTRACT.md, tooling/adapter_manifest.json, and ADAPTER_REGISTRY.json.
• [ ] For deployments that rely on TTLs or long-running memory usage, operators have an explicit plan or runbook entry to invoke memory.prune periodically as part of maintenance (no built-in scheduler is provided).

7) Final Publish Pass

• [ ] Run a final docs link check (manual or scripted).
• [ ] Verify no private/sensitive data in committed artifacts.
• [ ] Release Gates workflow is green (wheel import smoke, pip check, MCP dry-run gates).
• [ ] Wheel install check passed for import runtime.compat, adapters, cli.main.
• [ ] ainl install-mcp --host openclaw --dry-run and --host zeroclaw --dry-run both pass.
• [ ] Release notes include a clear install-regression status line (what was validated, and any remaining host-lane caveats).
• [ ] Create release notes summarizing:
  • key technical changes
  • known limitations
  • immediate next milestones