Eve library

Phase 130 completion evidence

Tracker

Revision 1

On this page

Historical task source. The Eve task board holds current task state.

Verified on 2026-07-14 in /mnt/volume300/workspace/oshun-lt-2.

Completion audit#

  • TODOS/phase-130.md contains 60 completed tasks and no unchecked task. The implementation maps to the seven Phase 130 libraries experiment-orchestration, doe, protocols, instrument-control, sample-ops, lab-safety, and closed-loop-experiments, plus the Phase 128 materials-graph integration required by section 130.8.
  • A source scan found no actionable TODO/FIXME markers, unimplemented stubs, or disabled tests in the phase libraries. The phase-specific TypeScript suites contain 54 tests: 51 across the seven Phase 130 libraries and three experimental-knowledge-capture tests in materials-graph.
  • The tests exercise every numbered implementation group: operating-model types and integration links; classical, constrained, Bayesian, and sequential DOE; planning outputs; protocol compilation, simulation, execution, recovery, evidence, and backend connectors; sample inventory, geometry, custody, batching, contamination, and lineage audits; instrument drivers, adapters, facility sessions, calibration, QC, and streaming; safety validation, access gating, review, incidents, and audit export; closed-loop orchestration, model updating, discrepancy learning, four domain loops, and convergence; and knowledge-graph fusion, reconciliation, and experiment packs.

Integration and packaging corrections#

  • Phase 128 stack links now identify the actual Phase 128 libraries: @kalika/materials-notebooks, @kalika/materials-collab, @kalika/discovery-agents, and @kalika/materials-graph. The orchestration project declares matching Nx dependency edges.
  • Experimental output ingestion, theory/experiment reconciliation, and publication/patent pack creation now live in and export from the Phase 128 materials graph instead of the unrelated general mathematical knowledge graph.
  • All seven Phase 130 package manifests now expose their real TypeScript source entrypoints. The four packages with cross-package imports declare the corresponding workspace:* dependencies, and the lockfile importer entries match those manifests.
  • All seven deprecated @nx/vite:test targets were migrated to @nx/vitest:test. Each library now has a strict type-check target, and the orchestration project exposes its Rust and Python tests as Nx targets.

Verification#

  • TypeScript: strict type-checks passed for all seven Phase 130 libraries and materials-graph; ESLint passed across all eight affected libraries.
  • Tests: all 51 tests in the seven Phase 130 TypeScript suites passed; the full materials-graph suite passed 16 tests, including all three section 130.8 tests.
  • Builds: targeted Nx builds passed serially for all eight affected libraries with dependency builds excluded so the evidence measures the changed package surfaces directly.
  • Native and adapter surfaces: Rust unit/doc tests passed (one unit test), cargo fmt --check passed, and Clippy passed with warnings denied. Python pytest passed (one adapter-contract test), with Ruff check/format and strict mypy also passing.
  • Package/lock validation: each new workspace dependency has a matching lockfile importer entry. A repository-wide frozen-lockfile install is not a valid Phase 130 gate in this worktree because the unrelated libs/neith/colorist-panel manifest already has an unstaged dependency absent from its lock importer.
  • The pre-commit affected-typecheck heuristic reaches unrelated reverse consumers and reports existing missing tslib, @types/three, and Node-type declarations in Nous V2, Kalika web, and Sophia. Those failures reproduce outside the Phase 130 packages and do not contradict the eight passing direct type-check targets above.
  • No Phase 130 library contains TSX/JSX or a mounted browser/mobile application surface. These packages provide orchestration and domain APIs, so Playwright, accessibility, and mobile UI automation are not applicable to this phase.

Resource-safe execution#

The host had 15 GiB RAM, 31 GiB swap, 44 GiB free disk, and approximately 13 GiB available memory before verification. Tests, type-checks, linting, builds, Rust, and Python gates were run with targeted Phase 130 scope; the eight-library Nx build used --parallel=1 and was supervised at 30-second intervals. Available memory remained at least 11 GiB during the most expensive command.