mirai-graph 1.4.0 → 2.1.0

package/releases/0.1.0-alpha.4.md

@@ -1,53 +0,0 @@
-# Mirai Graph v0.1.0-alpha.4
-
-Status: release notes
-
-## Theme
-
-Executable standard hardening plus the first public AI employee profile.
-
-## Included Since v0.1.0-alpha.3
-
-- Context-pack selection explanations with object relevance scores and relation
-  inclusion reasons.
-- Context-pack validation mode for generated pack metadata, graph references
-  and explanation coverage.
-- Negative fixture for missing context-pack selection explanations.
-- Positive profile conformance result artifacts.
-- Public capability map from practical graph-method areas to Mirai Graph
-  standard, schema, tooling, profile and research artifacts.
-- Alpha `ai_employee` profile.
-- Graph-as-AI-brain standard draft.
-- Action/runtime boundary standard draft.
-- Feedback learning gate standard draft.
-- Minimal synthetic AI employee fixture.
-- AI employee adoption guide.
-
-## Validation
-
-Validated with:
-
-```bash
-npm run release:check
-```
-
-The release also passed the public safety scan used by maintainers to check for
-private paths, obvious secrets and unsupported claims.
-
-## Claim Boundary
-
-This release should claim only:
-
-```text
-Mirai Graph v0.1.0-alpha.4 improves executable alpha validation and introduces
-the first public AI employee graph profile with explicit action/runtime and
-feedback-learning boundaries.
-```
-
-It should not claim:
-
-- real-world AI employee safety;
-- autonomous execution readiness;
-- Larena/AI runtime conformance;
-- productivity improvement;
-- external empirical validation.

package/releases/0.1.0-alpha.5.md

@@ -1,48 +0,0 @@
-# Mirai Graph v0.1.0-alpha.5
-
-Status: release notes
-
-## Theme
-
-Public standardization of graph-managed skill runtime patterns.
-
-## Included Since v0.1.0-alpha.4
-
-- Hybrid Source Of Truth standard draft.
-- Mirai Graph Adoption Levels standard draft with `GGA0..GGA10`.
-- Runtime Enablement standard draft for task-scoped graph-mode use.
-- Projection Views standard draft.
-- Alpha `skill_runtime` profile.
-- Minimal synthetic skill-runtime fixture.
-- Profile conformance result for the skill-runtime fixture.
-- Public-safe transfer note for reusable lessons from graph-managed skills.
-- README and profile fixture documentation updates.
-
-## Validation
-
-Validated with:
-
-```bash
-npm run release:check
-```
-
-The release also passed the public safety scan used by maintainers to check for
-private paths, obvious secrets and unsupported claims.
-
-## Claim Boundary
-
-This release should claim only:
-
-```text
-Mirai Graph v0.1.0-alpha.5 adds public alpha standard sections and an executable
-skill-runtime profile for governed capability graphs.
-```
-
-It should not claim:
-
-- graph-only runtime readiness;
-- semantic completeness;
-- operational effectiveness;
-- runtime safety;
-- universal validity;
-- publication of private skill internals.

package/releases/0.1.0-alpha.6.md

@@ -1,96 +0,0 @@
-# Mirai Graph v0.1.0-alpha.6
-
-Status: release notes
-
-## Theme
-
-Public standardization of executable Mirai Graph governance patterns.
-
-This alpha started with explainable federation routing control-loop patterns and
-now also packages the implementation-control and process-control layer needed
-to make Mirai Graph adoption inspectable with public-safe schemas, fixtures and
-validators.
-
-## Included Since v0.1.0-alpha.5
-
-### Routing Control
-
-- Route Explanation standard draft.
-- Routing Fixtures standard draft with forbidden route shapes.
-- Federation Health standard draft with conflict details.
-- Route Regression standard draft.
-- Public schemas for route explanation, routing fixtures, routing fixture run,
-  federation health dashboard and route regression result.
-- Synthetic federation routing smoke example with per-fixture route evidence,
-  high-risk explanation, route regression and federation health dashboard.
-- `validate:routing-control` validation command.
-- Public-safe transfer note for federation routing control-loop lessons.
-
-### Implementation And Process Control
-
-- Implementation Control profile hardening.
-- Implementation Control Cycles standard and synthetic example.
-- Launch Record standard, schema, example and validator.
-- Process Transition request schema, passing fixtures and negative fixtures.
-- Process Control Contract standard, schema, example and validator.
-- Work State Machine and Recovery/Resume schemas and examples.
-- Terminal Kaizen closure rule for significant process transitions.
-- Clear boundary between human/projection stages and executable state-machine
-  movement.
-
-### Governance Standard Modules
-
-- Graph DNA Alignment standard, schema and synthetic example.
-- Risk Control Matrix standard, schema and synthetic example.
-- Multi-Agent Coordination standard, schema and synthetic example.
-- Source Boundary Contract standard, schema and synthetic example.
-- Organization Governance profile, standard draft and synthetic example.
-- Public-safe project-management transfer note for implementation-control
-  patterns.
-
-### Developer Tooling
-
-- Validation commands for implementation-control cycles, launch records,
-  process transitions, process-control contracts, DNA alignment, work state,
-  recovery, risk, coordination, source-boundary and organization-governance
-  examples.
-- Negative test coverage for invalid graph packages and invalid process
-  transitions.
-
-### Navigation And Evidence Packaging
-
-- Role-based README navigation.
-- End-to-end control-loop tutorial.
-- Playground index for runnable CLI demos.
-- Scientific evidence package framing synthetic evidence, pilot evidence and
-  scientific claim boundaries.
-
-## Validation
-
-Validate with:
-
-```bash
-npm run release:check
-```
-
-## Claim Boundary
-
-This release should claim only:
-
-```text
-Mirai Graph v0.1.0-alpha.6 adds public alpha standard and schema shapes for
-explainable federation routing, implementation-control cycles, launch records,
-process transitions, process-control contracts, governance boundaries and
-public-safe evidence packaging.
-```
-
-It should not claim:
-
-- private federation behavior;
-- graph-only runtime readiness;
-- runtime safety;
-- owner-source rewrite permission;
-- scientific proof of effectiveness;
-- production readiness for autonomous execution;
-- automatic canonical update from generated context, feedback, evidence or
-  proposals.

package/releases/0.1.0-alpha.7.md

@@ -1,114 +0,0 @@
-# Mirai Graph v0.1.0-alpha.7
-
-Status: release preparation notes
-
-## Theme
-
-Readiness jump for explainable executable process control.
-
-This alpha makes Mirai Graph easier to inspect without maintainer guidance by
-adding process-transition decision explanations, a deterministic local
-playground report, baseline-comparison evidence and an additional public-safe
-pilot outside the original software and AI-employee examples.
-
-## Included Since v0.1.0-alpha.6
-
-### Process-Transition Explanations
-
-- Process-transition validator JSON output now includes an `explanation`
-  section.
-- Explanation output records current and target states, matched transition,
-  required and provided evidence, missing evidence, launch-record status,
-  terminal/Kaizen decision and false-transition guard summary.
-- Markdown process-transition output is now a transition decision report.
-- Passing and failing explanation fixtures are committed under
-  `examples/process-transition-minimal/results/`.
-- `validate:process-transition-report` checks explanation-critical fields.
-
-### Playground Report
-
-- Playground remains CLI-first and deterministic.
-- `playground:report` generates the same report as `playground:demo`.
-- `playground/demo-report.md` now shows:
-  - seed;
-  - graph package;
-  - context pack;
-  - launch record;
-  - transition explanation;
-  - evidence;
-  - Kaizen;
-  - baseline comparison.
-- `validate:playground-report` keeps the stored demo report aligned with the
-  generated output.
-
-### Baseline Comparison Evidence
-
-- Added `schemas/baseline-comparison-result.schema.json`.
-- Added public-safe synthetic result artifact at
-  `benchmarks/synthetic-context-reduction-v0/results/baseline-comparison-result.json`.
-- Added `validate:baseline-comparison`.
-- Evidence docs now separate synthetic comparison, pilot evidence and
-  scientific proof boundaries.
-
-### Public-Safe Pilot Coverage
-
-- Added independent implementation pilot 004:
-  `pilots/independent-implementation-004-research-program/`.
-- The pilot models a synthetic research-program governance workflow using the
-  existing `project_management` profile.
-- The pilot includes seed, package, graph objects, graph relations, gates,
-  report and profile-conformance result.
-- Public safety boundary excludes private research repositories, reviewer
-  notes, handoff text, internal traces, secrets and customer or employee data.
-
-### Release Packaging
-
-- Updated package version to `0.1.0-alpha.7`.
-- Updated README navigation, roadmap, playground docs, evidence docs,
-  conformance fixture docs and changelog.
-
-## Validation
-
-Validate with:
-
-```bash
-npm test
-npm run release:check
-```
-
-Targeted alpha.7 checks:
-
-```bash
-npm run validate:process-transition-report
-npm run playground:report
-npm run validate:playground-report
-npm run validate:baseline-comparison
-npm run validate:pilots
-```
-
-## Claim Boundary
-
-This release should claim only:
-
-```text
-Mirai Graph v0.1.0-alpha.7 adds public alpha readiness for explainable
-process-transition reports, deterministic local demos, bounded baseline
-comparison evidence and wider public-safe pilot coverage.
-```
-
-It should not claim:
-
-- peer-reviewed proof of effectiveness;
-- production runtime safety;
-- autonomous execution readiness;
-- graph-only runtime readiness;
-- universal superiority over document or task-tracker workflows;
-- automatic canonical update from generated context, feedback, evidence,
-  runtime results or proposals.
-
-## Known Limits
-
-- CLI remains reference tooling, not a full SDK.
-- Playground is deterministic and local, not a hosted interactive environment.
-- Baseline comparison is synthetic and requires replication.
-- Public pilots show applicability, not broad external validity.

package/releases/0.1.0-alpha.8.md

@@ -1,92 +0,0 @@
-# Mirai Graph v0.1.0-alpha.8
-
-Status: release preparation notes
-
-## Theme
-
-Adopter kit and executable technology-quality control.
-
-This alpha keeps Mirai Graph CLI-first and public-safe while making adoption
-more practical through starter templates, organization-governance pilot
-coverage, a unified CLI wrapper and a technology quality feedback gate that
-checks whether work followed the declared process before acceptance or release.
-
-## Included Since v0.1.0-alpha.7
-
-### Adopter Kit
-
-- Added starter templates for software projects, research programs,
-  AI employees and organization governance.
-- Added unified `mirai_graph` CLI wrapper for common validation, report and
-  process-transition explanation commands.
-- Added organization-governance pilot coverage.
-- Added Evidence Package v0.2 as the current public research evidence
-  entrypoint.
-
-### Technology Quality Feedback
-
-- Added `standard/technology-quality-feedback.md`.
-- Added `schemas/technology-quality-feedback.schema.json`.
-- Added synthetic example:
-  `examples/technology-quality-feedback-minimal/`.
-- Added CLI mode:
-  `node packages/cli/validate-mirai-graph.js technology-quality-feedback <file>`.
-- Added `validate:technology-quality-feedback` and
-  `test:technology-quality-feedback-negative`.
-- Added fail-closed negative fixtures for blocking findings, missing process
-  steps, unclassified findings, missing Kaizen routes, unresolved spec gaps,
-  missing feedback reports and evidence-as-authorization errors.
-
-### Process-Control Integration
-
-- Process-control docs now state that significant acceptance, approval or
-  release transitions should require technology quality feedback.
-- The implementation-control profile now includes technology contracts,
-  technology quality feedback reports and quality findings.
-- The process-control contract schema supports an optional
-  `technology_quality_feedback_policy`.
-
-## Validation
-
-Validate with:
-
-```bash
-npm test
-npm run release:check
-```
-
-Targeted alpha.8 checks:
-
-```bash
-npm run validate:technology-quality-feedback
-npm run test:technology-quality-feedback-negative
-npm run validate:process-control
-npm run validate:cli-unified
-```
-
-## Claim Boundary
-
-This release should claim only:
-
-```text
-Mirai Graph v0.1.0-alpha.8 adds public alpha readiness for adopter templates,
-unified CLI entrypoints, organization-governance pilot coverage and executable
-technology-quality feedback gates.
-```
-
-It should not claim:
-
-- peer-reviewed proof of effectiveness;
-- production runtime safety;
-- autonomous execution readiness;
-- graph-only runtime readiness;
-- automatic acceptance from passing tests;
-- automatic canonical update from generated context, feedback, evidence,
-  runtime results or proposals.
-
-## Known Limits
-
-- CLI remains reference tooling, not a full SDK.
-- Technology quality feedback validates report shape and boundary rules; it
-  does not prove semantic correctness by itself.
-- Public examples remain synthetic and require external replication.

package/releases/0.1.0-alpha.9.md

@@ -1,80 +0,0 @@
-# Mirai Graph v0.1.0-alpha.9
-
-Status: published GitHub prerelease; npm publication pending registry
-authentication.
-
-## Theme
-
-Public naming consolidation under Mirai Graph.
-
-This alpha consolidates the public standard, package, repository and CLI
-identity around the `mirai-graph` slug.
-
-## Included Since v0.1.0-alpha.8
-
-### Rename Transition
-
-- Public standard name is now Mirai Graph.
-- Package name is now `mirai-graph`.
-- Primary CLI entrypoint is now `packages/cli/mirai-graph.js`.
-- Preferred package manifest is now `mirai-graph-package.json`.
-- The underscore CLI alias `packages/cli/mirai_graph.js` is available for local
-  environments that prefer identifier-style command names.
-
-### Command Surface
-
-- `packages/cli/mirai-graph.js` is the primary local CLI.
-- `packages/cli/mirai_graph.js` delegates to the primary CLI as an underscore
-  alias.
-- Validators use `mirai-graph-package.json`.
-- `npm test` includes a smoke check for the primary CLI, underscore alias and
-  current manifest filename.
-
-### Documentation
-
-- Added a transition guide:
-  `docs/adoption/mirai-graph-transition.md`.
-- Updated README, CLI docs, citation metadata, changelog and release navigation
-  for Mirai Graph naming.
-
-### Release State
-
-- GitHub repository: `zabarov/mirai-graph`.
-- GitHub release:
-  `https://github.com/zabarov/mirai-graph/releases/tag/v0.1.0-alpha.9`.
-- Git tag: `v0.1.0-alpha.9`.
-- NPM package: pending publication until npm registry authentication is
-  available.
-
-## Validation
-
-Validate with:
-
-```bash
-npm run validate:mirai-graph-compat
-npm test
-npm run release:check
-```
-
-## Claim Boundary
-
-This release should claim only:
-
-```text
-Mirai Graph v0.1.0-alpha.9 consolidates public naming, package identity and
-local CLI surfaces around Mirai Graph.
-```
-
-It should not claim:
-
-- schema namespace ownership has been finalized;
-- npm package publication has happened;
-- historical evidence payloads were regenerated;
-- peer-reviewed proof of effectiveness.
-
-## Known Limits
-
-- NPM publish remains pending registry authentication.
-- Schema `$id` namespace migration remains pending domain confirmation.
-- The underscore CLI alias is an implementation convenience, not a separate
-  public standard.

package/releases/1.0.0-rc.1.md

@@ -1,121 +0,0 @@
-# Mirai Graph v1.0.0-rc.1
-
-Status: release candidate prepared in repository; tag, GitHub Release and npm
-publish require separate release execution.
-
-## Theme
-
-1.0 Release Candidate.
-
-This release candidate consolidates the public Mirai Graph standard from the
-alpha line into a single verifiable package for external review. It does not
-change the core claim boundary: Mirai Graph validates graph-managed structure,
-evidence, process state and governance gates; it does not execute production
-actions or authorize canonical updates by itself.
-
-## Included Scope
-
-### Core Standard
-
-- Object model.
-- Relation model.
-- Lifecycle and readiness.
-- Evidence provenance.
-- Context-pack generation.
-- Governance and conformance levels.
-- Graph seed and graph embryo.
-- Hybrid source-of-truth boundary.
-
-### Domain Profiles
-
-- Software specification.
-- Project management.
-- AI employee.
-- Character Layer.
-- Skill runtime.
-- Implementation control.
-- Organization governance.
-
-### Executable Process Control
-
-- Implementation control cycles.
-- Launch records.
-- Work state machine.
-- Process-control contract.
-- Process-transition validation and readable explanations.
-- Technology quality feedback.
-- Recovery/resume.
-- Risk-control matrix.
-- Source-boundary contract.
-- Multi-agent coordination.
-
-### Instrumentation And Quality Feedback
-
-- Development cockpit.
-- Feature implementation traceability.
-- Multi-source quality feedback.
-- Instrumentation readable report.
-- Negative fixtures that fail closed on missing gates, missing evidence and
-  blocked acceptance.
-
-### Character Layer
-
-- Public `character_layer` profile.
-- Character Layer starter pack.
-- Semantic anti-fixtures.
-- Cross-layer integration artifact for AI employee, process control,
-  technology quality feedback and runtime context.
-- Model-portability evidence shape.
-- Character Layer readiness artifact for the Mirai Graph 1.0 track.
-
-### Adoption And Reproducibility
-
-- Self-service onboarding commands: `detect`, `bootstrap` and `init`.
-- GitHub Action starter for package validation.
-- "Connect A Project In 15 Minutes" tutorial.
-- Role-based adopter kit.
-- Starter templates for developer, researcher, AI employee, behavior governance
-  and organization governance paths.
-- Deterministic playground report.
-- Synthetic baseline comparison.
-- Public-safe pilot packages.
-- Evidence package v0.3 and bounded scientific claim framing.
-
-### Proposal / Experimental Surface
-
-- Societal Governance profile proposal and synthetic fixtures are included as a
-  proposal-level public-safe exploration surface. They are not core
-  `1.0.0-rc.1` conformance requirements.
-
-## Validation
-
-```bash
-npm test
-npm run release:check
-npm run validate:release-state
-npm run validate:character-layer-readiness
-npm run validate:instrumentation-layer
-npm run validate:technology-quality-feedback
-npm run validate:baseline-comparison
-```
-
-## Claim Boundary
-
-This release candidate supports the claim that Mirai Graph has a coherent,
-validator-backed public standard shape for graph-managed knowledge, process,
-AI-employee behavior, implementation control, instrumentation and evidence.
-
-It does not claim:
-
-- peer-reviewed scientific proof;
-- broad external validity;
-- production autonomous execution safety;
-- model-backend equivalence;
-- automatic canonical updates from generated context, evidence, feedback,
-  cockpit metrics or proposals;
-- npm package publication until npm release-state checks confirm publication.
-
-## Release Execution Boundary
-
-This file is release preparation only. Creating a git tag, GitHub Release or npm
-publication must happen as a separate explicit release execution step.

package/releases/1.0.0-rc.10.md

@@ -1,14 +0,0 @@
-# Mirai Graph 1.0.0-rc.10
-
-This release candidate completes the public Project Technology pilot against
-the real Federation graph.
-
-## Verified outcome
-
-- commands without a repository argument use the caller's project;
-- large tracked-file inventories are read completely;
-- accepted target objects cannot disappear because of a process-output limit;
-- the Project Technology contract and fail-closed safety rules are unchanged.
-
-Stable `1.0.0` remains deferred until this RC has been exercised by the SIMAI
-Federation and all release-managed skills.

package/releases/1.0.0-rc.11.md

@@ -1,14 +0,0 @@
-# Mirai Graph 1.0.0-rc.11
-
-This release candidate closes the final architecture-owner gap found by the
-real Federation process resolver.
-
-## Verified outcome
-
-- complete large-repository target discovery;
-- exact human architecture owner in the bounded contract;
-- component, package, capability and operational ownership remain digest-bound;
-- no private source or full provider graph is exported.
-
-Stable `1.0.0` remains deferred until Federation 8.20.0 and all managed skills
-complete their release checks.