Prepare aoa-techniques v0.5.0 release

CHANGELOG.md

@@ -6,61 +6,145 @@ The format is intentionally simple and human-first.
 
 ## [Unreleased]
 
+## [0.5.0] - 2026-07-13
+
+### Summary
+
+- this release reconstructs the complete post-`v0.4.5` delta from Git history
+  and source diffs: `37` first-parent commits across `366` paths, with `98,241`
+  additions and `22,697` deletions; `31` of those commits did not edit this
+  changelog and are therefore accounted for explicitly below
+- validation now has manifest-owned source-fast, generated, part-local,
+  release, nightly, and advisory lanes, with owner-split validators, test and
+  script topology inventories, dedicated GitHub workflows, and a stabilizing
+  release entrypoint
+- the repository now carries bounded local memo, eval, KAG, and stats ports
+  while preserving `aoa-memo`, `aoa-evals`, `aoa-kag`, and `aoa-stats` as the
+  stronger owners of their respective cross-repository semantics
+- KAG export and repo-local indexes now carry artifact identity, trust-root and
+  subject-store admission, canonical index-family parity, squash-stable event
+  history, and fail-closed consumer verdicts
+- this remains a public technique canon and generated companion surface, not a
+  skill runtime, proof authority, routing engine, memory owner, KAG substrate,
+  cross-owner stats layer, or package-registry artifact
+
+### Added
+
+- added the root `memo/` port, portable validation route, reviewed-memory route
+  labels, and memory-trigger guidance without turning session evidence into
+  technique truth
+- added canonical decision indexes plus modeled-surface contracts that reject
+  implicit, unlisted, or unmodeled decision-lane inputs
+- added the manifest-backed validation topology, source-only PR gate,
+  generated projection groups, part-local discovery runner, release/nightly
+  workflows, and explicit test and script inventories
+- added the root `evals/` skeleton for technique-local eval pressure while
+  keeping verdict, scoring, regression, and proof doctrine with `aoa-evals`
+- added artifact identity and OS Abyss trust-envelope validation for the KAG
+  export, including durable evidence promotion and subject-store admission
+- added the root `kag/` provider home, source-return packet, provider readiness
+  surface, repo-local indexes, canonical seven-index family, and downstream
+  parity validation
+- added the root `stats/` owner port with a revision-bound
+  promotion-readiness pass ratio, central protocol validation, and coverage
+  that rejects a false numerator while preserving technique and audit authority
+
 ### Changed
 
-- added the root `stats/` owner port with a revision-bound promotion-readiness
-  pass ratio, central protocol validation, and coverage that rejects a false
-  numerator while preserving technique and audit authority
-- added the root `kag/` local provider district and wired it into root surface
-  law, the AGENTS mesh, and generated mesh index for downstream KAG consumers
-- tightened the OS Abyss KAG export artifact envelope so the manifest, validator,
-  and downstream contract test require durable evidence promotion,
-  materialized subject-store admission, source/trust-root matching, and
-  fail-closed allow/deny consumer verdicts before agent consumption
-- moved validation command authority into `config/validation_lanes.json`, added
-  `scripts/validation_lanes.py` and `scripts/ci_gate.py`, and changed GitHub
-  `Repo Validation` to run the `source-fast` lane instead of the full
-  `scripts/release_check.py` release gate on every PR
-- added a validation topology district with source-fast, generated,
-  mechanics/part-local, release, nightly, and advisory lane definitions scoped
-  to the technique-canon owner boundary
-- split GitHub release and nightly validation into separate pinned-action
-  workflows so pull requests do not run the full release-freeze gate
-- split `scripts/validate_repo.py` into `scripts/validators/` owner modules
-  while keeping the old CLI/import surface as a thin compatibility adapter
-- isolated generated/read-model checks into explicit projection validator
-  modules and generated lane groups for catalog, decisions, AGENTS mesh,
-  mechanics projections, KAG export, Technique Intelligence, questbook, and
-  public hygiene
-- added `docs/testing/TEST_TOPOLOGY.md`, `docs/testing/test_inventory.json`,
-  and topology tests so root, mechanic-level, part-local, and Spark agent-lane
-  test homes stay inventoried without becoming release command authority
-- changed the `mechanics/part-local` lane to enter through
-  `scripts/run_part_local_tests.py`, which discovers part-local pytest files
-  and their related builder `--check` / validator scripts from source homes
-- split the former `tests/test_validate_repo.py` monolith into owner-surface
-  tests for compatibility imports, CI/release authority, source contracts,
-  generated drift, AGENTS mesh, public hygiene, and questbook/intelligence
-- split the former Distillation mechanics topology monolith into phase, part,
-  and surface-owned suites with separate inventory failure routes
-- changed active docs and decision command guidance to name validation lane ids
-  and nearest owner `AGENTS.md` checks instead of copied release snippets
-- expanded `source-fast` with a source-only
-  `scripts/validate_source_contracts.py` entrypoint so PR validation covers
-  authored technique contracts without pulling generated freshness into the
-  growth gate
-- added guardrail coverage for active `AGENTS.md` cards so deleted test targets
-  such as the former validate-repo and Distillation topology monoliths cannot
-  remain as focused local checks
-- expanded the nightly workflow with a latest `v*` release reproduction job
-  while keeping moving-main drift checks separate from release-freeze behavior
-- added script topology coverage with `docs/validation/SCRIPT_TOPOLOGY.md`,
-  `docs/validation/script_inventory.json`, and `tests/test_script_topology.py`
-  so every active `*/scripts/*` surface has an explicit owner, lane,
-  side-effect boundary, CI inclusion, and focused test route
-- added a root-level `evals/` skeleton port for technique-local eval pressure
-  while central verdict, scoring, regression, and proof doctrine stay owned by
-  `aoa-evals`
+- changed root and generated agent guidance to use the canonical `aoa-memo`
+  route label and explicit memory-route trigger law
+- moved executable validation command authority into
+  `config/validation_lanes.json`; active docs and decisions now name lane ids
+  and nearest owner cards rather than copying release command sequences
+- split the former repo validator and its monolithic test into owner-local
+  source, generated, AGENTS-mesh, hygiene, questbook, intelligence, and
+  compatibility modules, and similarly split Distillation topology coverage
+  into phase-, part-, and surface-owned suites
+- separated ordinary PR growth validation from release-freeze and nightly
+  reproduction, while keeping latest-tag reproduction and generated drift
+  checks explicit
+- keyed live recurrence receipts by event id and sanitized owner-local evidence
+  and workspace-origin paths before they enter public technique surfaces
+- aligned KAG provider validation with the canonical shared contract, then
+  adopted pinned, lineage-aware, streamable, squash-stable repository index
+  generation and event-history propagation
+- tightened KAG export admission so materialized subjects, source and trust-root
+  identity, evidence state, and allow/deny consumer verdicts must agree before
+  agent consumption
+
+### Fixed
+
+- restored Spark center-decision references and made the local memo validator
+  route portable
+- enforced CSV parity for the kind overlay and repaired the Distillation reform
+  context guard, Wave A anatomy counts, and promotion-evidence hold label
+- repaired owner-truth closeout and technique-tree relative links across source
+  bundles and their generated companions
+- enforced bundle-review draft kind, hardened antifragility and recurrence
+  owner-request receipt tests, and retained wrapped rubric finding tokens
+- aligned the published-summary remediation snapshot cap policy with its
+  checklist, generated readers, and regression coverage
+
+### Included changes
+
+Every first-parent commit from `v0.4.5` through the pre-release base
+`e45b41f5` is listed exactly once, in landing order:
+
+- `5d73dfa3` — `[codex] Add memory route trigger law (#457)`
+- `c030988e` — `Use canonical aoa-memo route label (#458)`
+- `9a0ac2ba` — `Add local memo port (#459)`
+- `b8aeb335` — `Canonicalize technique decision indexes (#460)`
+- `5a733764` — `[codex] Close validation lane coverage gaps (#461)`
+- `4e3ceb77` — `[codex] Add script topology coverage`
+- `3df26cb3` — `Detect unmodeled decision lane surfaces (#463)`
+- `a0263cd4` — `Honor modeled decision lane surfaces (#464)`
+- `01461d6b` — `Require modeled surfaces to be explicit lists (#465)`
+- `1a7d1469` — `Add local eval port skeleton`
+- `ad926b20` — `Restore Spark center decision references (#467)`
+- `7c327f0e` — `Require portable memo validation route (#468)`
+- `60e8edee` — `Validate kind overlay CSV parity (#469)`
+- `b008e207` — `Fix distillation reform context guard (#470)`
+- `1fdf90ab` — `Fix bundle anatomy Wave A counts`
+- `e30e26fa` — `Define promotion evidence hold rubric label`
+- `9c28aab7` — `Fix owner truth closeout links`
+- `0d56be3b` — `Fix stale technique tree relative links (#474)`
+- `6b51b6ea` — `Enforce bundle review draft kind (#475)`
+- `8b2ad2de` — `Harden antifragility ORQ receipt test (#476)`
+- `1fdf9849` — `Harden recurrence ORQ receipt test (#477)`
+- `30a70271` — `Capture wrapped rubric finding tokens (#478)`
+- `15bafd85` — `[codex] Add artifact identity to KAG export (#479)`
+- `5fc1c0c2` — `Add OS Abyss trust gate for KAG export (#480)`
+- `b5e38e08` — `Require subject-store trust gate for KAG export (#481)`
+- `8ab16bc8` — `Add local KAG provider home (#482)`
+- `50f14c92` — `Align KAG provider validation route (#483)`
+- `326f70bf` — `Add repo-local KAG indexes (#484)`
+- `89432bbc` — `Key live receipt hooks by event id (#485)`
+- `a1918b54` — `Sanitize promotion incubation evidence handles (#486)`
+- `e432adac` — `Sanitize workspace ingress origin path (#487)`
+- `98a5533f` — `Align remediation snapshot cap policy (#488)`
+- `67f90cc7` — `Enforce repo-local KAG index parity (#489)`
+- `d58fd66c` — `Adopt repository KAG index family (#490)`
+- `5fa63ab7` — `Publish canonical repository KAG indexes (#491)`
+- `090440b5` — `Add aoa-techniques stats port (#492)`
+- `e45b41f5` — `Harden KAG export artifact admission (#493)`
+
+### Validation
+
+- exact `37`-of-`37` first-parent inventory parity against the tagged range
+- complete `366`-path source-diff accounting, including the `31` commits that
+  did not edit the prior changelog
+- source-fast, generated, mechanics/part-local, and release validation lanes
+- repository tests, public-hygiene checks, and diff hygiene
+- repo-local KAG full, incremental, family, and event-history parity
+- landed-main GitHub validation and post-merge release-lane reproduction
+
+### Notes
+
+- generated outputs remain derived companions; authored technique bundles,
+  route cards, contracts, decisions, mechanic-owned packets, and local port
+  sources remain authoritative
+- package publication to PyPI, npm, or another registry remains out of scope;
+  the public artifact is the tagged corpus and its GitHub release
 
 ## [0.4.5] - 2026-05-18
 

README.md

@@ -11,7 +11,7 @@ This README is only the public front door. When the question becomes authority,
 authoring, classification, review, release, mechanics, or agent routing, follow
 the linked owner surface instead of expanding root prose.
 
-> Current release: `v0.4.5`. See [CHANGELOG](CHANGELOG.md) for release notes.
+> Current release: `v0.5.0`. See [CHANGELOG](CHANGELOG.md) for release notes.
 
 ## What This Repository Does
 

docs/readers/repo/REPO_DOC_SURFACES.md

@@ -73,7 +73,7 @@ Use these when the question is what changed, what is currently unreleased, and h
 |---|---|---|
 | [Roadmap](../../../ROADMAP.md) (`ROADMAP.md`) | live repo-level direction and horizon posture for technique-canon growth | `Authority`, `Update Rule`, `Current Direction`, `Current Checked Contour`, `Horizon: Root Clarity`, `Horizon: Technique Atom`, `Horizon: Corpus Topology`, `Horizon: Corpus Tree`, `Horizon: Small-Agent Usability`, `Horizon: Mechanics To Canon`, `Horizon: Evidence And Promotion`, `Horizon: Standalone Portability`, `Horizon: Generated Companions`, `When The Time Comes`, `Standing Direction` |
 | [QUESTBOOK.md — aoa-techniques](../../../QUESTBOOK.md) (`QUESTBOOK.md`) | compact root index for durable technique-canon obligations that should survive the current diff | `Update trigger`, `Frontier`, `Near`, `Latent / parked`, `Harvest candidates`, `Backing files`, `Rule` |
-| [Changelog](../../../CHANGELOG.md) (`CHANGELOG.md`) | release and unreleased status history for the public corpus | `[Unreleased]`, `[0.4.5] - 2026-05-18`, `[0.4.2] - 2026-04-23`, `[0.4.1] - 2026-04-19`, `[0.4.0] - 2026-04-10`, `[0.3.0] - 2026-04-01`, `[0.2.0] - 2026-03-23`, `[0.1.0] - 2026-03-17` |
+| [Changelog](../../../CHANGELOG.md) (`CHANGELOG.md`) | release and unreleased status history for the public corpus | `[Unreleased]`, `[0.5.0] - 2026-07-13`, `[0.4.5] - 2026-05-18`, `[0.4.2] - 2026-04-23`, `[0.4.1] - 2026-04-19`, `[0.4.0] - 2026-04-10`, `[0.3.0] - 2026-04-01`, `[0.2.0] - 2026-03-23`, `[0.1.0] - 2026-03-17` |
 | [Releasing `aoa-techniques`](../../RELEASING.md) (`docs/RELEASING.md`) | bounded release flow and validation path for public docs and technique updates | `Release goals`, `Recommended release flow`, `Generated surface checklist`, `Release note shape`, `Versioning guidance`, `What not to optimize yet`, `Current stance` |
 
 ## Boundaries

generated/repo_doc_surface_manifest.json

@@ -350,6 +350,7 @@
       "bounded_role": "release and unreleased status history for the public corpus",
       "top_level_sections": [
         "[Unreleased]",
+        "[0.5.0] - 2026-07-13",
         "[0.4.5] - 2026-05-18",
         "[0.4.2] - 2026-04-23",
         "[0.4.1] - 2026-04-19",