agentera 3.0.0-dev.20 → 3.0.0-dev.22

package/README.md

@@ -48,6 +48,25 @@ It can remove only the exact Agentera-owned legacy link. Historical transcript
 import is also explicit (`agentera report refresh --import-source claude`) and
 is excluded from default active-runtime analytics.
 
+## Bounded state retrieval
+
+The executable contract is
+[`references/artifacts/state-storage-authority.yaml`](../../references/artifacts/state-storage-authority.yaml)
+and is projected by `agentera schema --format json`.
+
+```bash
+agentera state plan list --format json
+agentera state plan get --plan PLAN_ID --format json
+agentera state plan tasks list --limit 20 --format json
+agentera state experiments list --objective OBJECTIVE_ID --format json
+agentera state experiments get --objective OBJECTIVE_ID --number N --format json
+```
+
+Pages use opaque snapshot cursors, explicit omission fields, whole-entry output
+bounds, and exact retrieval. Plan history is owned by active/archive plan files;
+plan task retrieval is active-only. Experiment history is objective-scoped and
+reports full, summary-only, or unavailable detail without fabricating archives.
+
 See [UPGRADE.md](../../UPGRADE.md) for ownership, recovery, and migration
 details and [runtime feature parity](../../references/adapters/runtime-feature-parity.md)
 for host-specific behavior.

package/bundle/CHANGELOG.md

@@ -2,13 +2,53 @@
 
 ## [Unreleased]
 
+### Changed
+
+- Eliminated the monolithic `corpus.json` publication output from the extract CLI; report dry-run diagnostics now reference tier files and the `privacy.corpus_write` field is renamed to `privacy.tier_write`.
+- Changed `agentera check lint` verbosity checks to consume schema-owned full-file and per-entry word limits, skip no-limit and token-only dimensions, and report authority failures without fallback budgets.
+- Corrected packaged experiment-publication authority to declare validated objective-scoped archive-before-projection ownership and reject contradictory structured scope metadata.
+- Synchronized bounded plan, plan-task, and experiment retrieval across help, schema introspection, capability guidance, operator docs, packaged contracts, output bounds, archive ownership, and structured recovery errors.
+
 ### Added
 
+- Added an authoritative evidence-tier contract (`references/analysis/evidence-tier-authority.yaml`) and contract-model loader (`packages/cli/src/registries/evidenceTierContract.ts`) fixing the bounded full-evidence and signal tiers, supported Codex, Cursor, OpenCode, Copilot, and historical Claude source coverage, the consumer field-supply map including the latent startup-analysis reader, reserved signal semantics for the deferred glossary, and deterministic recovery outcomes for oversized, legacy, missing, corrupt, and incomplete state.
+- Added bounded tier publication and direct retrieval (`packages/cli/src/analytics/extractCorpus/evidenceTiers.ts`): complete local evidence is retained as independently bounded, source-family shards that split when exceeding the reader cap; a derived, content-addressable bounded signal tier carries only contract-required fields plus a resolvable `evidence_anchor`; generations are staged and revealed by an atomic current-pointer swap so no consumer observes a partial publication; and direct retrieval resolves a signal identity to its full record. The monolithic `corpus.json` write is eliminated; tiers are the only canonical publication output.
+- Added bounded tier-aware reads for usage analytics, report status, prime corpus coverage, startup analysis, and profile synthesis so every analytics consumer reads from bounded evidence shards one at a time instead of loading the monolithic `corpus.json`. A new `tierReader.ts` module projects contract compatibility states and recovery guidance, streams full-evidence records, and reads signal-tier metadata. A new `profileSignals.ts` module reads the bounded signal tier, assesses per-family retention sufficiency against the contract threshold, and resolves synthesized claims' evidence anchors to retained full-evidence shards. Legacy `corpus.json` remains a fallback when no tiers are published; oversized legacy state is never loaded whole.
+- Added a schema-owned verbosity-budget authority that classifies numeric word limits, explicit no-limit declarations, and token-only budgets while rejecting malformed, ambiguous, unreadable, and unsupported ownership without fallback limits.
+- Added a `profile_sufficiency` section to the evidence-tier authority defining `minimum_family_retention=0.5`, profile signal types (decision, question, correction, instruction, configuration), and insufficiency rules so bounded profile synthesis surfaces underrepresented families without fabricating confidence.
+- Added `agentera state experiments list|get --objective OBJECTIVE_ID` with newest-first bounded pages, snapshot-bound opaque cursors, merged projection/archive identity, verified archive-first detail, explicit provenance and detail availability, and structured compatibility failures.
+- Added objective-scoped immutable full-detail archives under `.agentera/optimize/<objective>/archive/experiments/<number>.yaml` before 10/40/50 projection replacement, with explicit provenance, conflict rejection, and crash-consistent idempotent retries.
+- Added `agentera state experiments publish --objective OBJECTIVE_ID --number N --input EXPERIMENT.yaml` with schema and identity preflight, atomic projection replacement, idempotent interruption retry, collision safety, and existing 10/40/50 projection behavior.
+- Added stable objective UUID and objective-scoped experiment identity validation, including experiment `0`, rename continuity, canonical `optimize` plus legacy `optimera` discovery, readable missing/duplicate legacy entries, and effect-free ambiguity rejection for future publication.
+- Added `agentera state plan list` and `agentera state plan get --plan PLAN_ID` for deterministic, cursor-paginated active and archived plan retrieval, full-plan exact reads, lifecycle provenance, and explicit legacy, invalid, missing, and ambiguous archive outcomes.
+- Added persisted plan identity across active-to-archive transitions, deterministic read-only identity for legacy plans, explicit lifecycle provenance, mirrored-path reporting, and structured ambiguity failures for identity collisions.
+- Added `agentera state plan tasks list` and `agentera state plan tasks get --task N` for bounded active-plan task retrieval with declared task-number order, opaque snapshot cursors, structured not-found and compatibility errors, and exact omission recovery in text and JSON plan output.
+
+- Added an executable bounded-retrieval authority and reserved CLI grammar for plan tasks, plans, and objective-scoped experiments, including stable and legacy identity rules, experiment `0`, opaque cursors, explicit omissions, and structured compatibility errors; retrieval execution remains staged for subsequent tasks.
 - Added a canonical runtime lifecycle projection with deterministic snapshot identity, applicability, ownership, action classes, counts, and command eligibility for OpenCode, Codex, Cursor, and Copilot.
 - Added bounded prime/status lifecycle attention rows with runtime-scoped repair previews, exact manual procedures, and doctor diagnostics for non-executable blockers.
+- Added runtime lifecycle schemas and adapters for skills, plugins, hooks, agents, configuration, enablement, trust, and native actions across all four active runtimes.
+- Added append-only ownership-journal recovery, exact operation dependencies, Linux secure directory-relative publication, and outcomes `applied`, `noop`, `failed`, `blocked_unowned`, `skipped_dependency`, and `action_required`.
+- Added explicit retired cleanup through `agentera upgrade --legacy-cleanup claude --dry-run|--yes`; it can remove only the ledger-owned legacy Agentera skill link and excludes user data.
+- Added typed artifact writes under `agentera state`: progress/decision/audit appends, decision satisfaction updates, plan create/task/status/archive lifecycle operations, schema-derived `explain`, server-assigned identifiers, exact retry replay, dry-run diffs, project locking, strict mapped-path safety, staged 10/40/50 compaction, final-byte validation, and atomic publication.
+- Added `install_track` to `prime --format json` (`app_home.install_track`) and `doctor --format json`. Values are `v2`, `v3`, `source`, and `unknown`.
+- Added the native TypeScript CLI on the npm development channel (`npx -y agentera@next`), bundling app data at pack time via `prepack` (`tsc` + `scripts/copy-bundle.mjs`) and the `.agentera-npx-bundle.json` sentinel checked by `packages/cli/test/packaging/prepack.test.ts`.
+- Added `bash scripts/single-binary.sh` and the packaging contract in `docs/packaging/v3-packaging.md` for Bun `--compile` single-binary delivery alongside the npm tarball.
+- Added `packages/cli/src/migrate/v2HandoffManifest.ts` to read `v3-handoff.json` during v2→v3 cleanup preflight per `references/cli/v3-handoff-manifest.schema.yaml`.
+- Added v3 `agentera doctor` coexistence detection in `packages/cli/src/upgrade/coexistenceProbe.ts` with regression coverage in `packages/cli/test/cli/coexistenceProbe.test.ts`.
+- Added maintainer gates under `agentera check` (`validate`, `verify`, `lint`, `compact`) with JSON parity oracles for validate families, verify eval targets, invalid-input envelopes, and `source_contract` emissions on `prime`, `prime --context`, `state plan`, and `state query --list-artifacts`.
+- Added per-channel `next_major` metadata in `references/cli/update-channels.yaml` and a six-line successor advisory at the head of `agentera doctor` on the stable channel, sourced from `packages/cli/src/upgrade/nextMajorDoctor.ts`.
+- Added a 1000-line monolith lint regression in `packages/cli/test/lint/monolithLint.test.ts` after splitting the eight largest `packages/cli/src/**/*.ts` modules by responsibility.
 
 ### Changed
 
+- Changed `agentera state migrate` to inventory legacy project state locally within
+  bounded file and byte limits, retain custom names with explicit provenance, and
+  require exact selectors plus `--apply --force` before immutable archive, backup,
+  and projection publication.
+- Changed `agentera state backfill` to remain optional local-Git enrichment: direct
+  `--apply --force` revalidates the project, `HEAD`, candidate, reachability, and
+  immutable target, while bounded history and no-remote behavior remain explicit.
 - Changed `agentera upgrade --only artifacts` to preview and apply deterministic
   plan lifecycle migration for docs-mapped current plans and archives. Legacy
   `active` and `completed` plan statuses become `open` and `complete` without
@@ -27,88 +67,42 @@
 - Changed lifecycle repair to fail closed on incomplete ownership provenance, isolate narrowed runtime selectors, and count explicit Claude cleanup separately from active-runtime work.
 - Changed corrupt or disconnected ownership-journal blockers to report stable causes without volatile event filenames across preview and apply.
 - Synchronized lifecycle projections across prime, status, project integration, and doctor while keeping Cursor CLI required, IDE conditional, and Claude cleanup explicit.
-
-### Fixed
-
-- Fixed full-plan task guidance and startup metadata to use lifecycle and coherence boundaries instead of a fixed task-count ceiling; typed publication still validates task identity and dependencies.
-- Fixed `agentera state plan create` and `archive` publication to use stable archive identities across retries, validate final serialized bytes with strict prose lint and schema checks, and retain canonical `overall acceptance` vocabulary.
-- Fixed `agentera state plan`: `set-status --task` accepts task values, and `set-plan-status` accepts `open` or `complete`.
-- Fixed plan discovery to derive activity from canonical paths: archive-only history is non-executable and cannot trigger a Build completion sweep; next work is dependency-ready, lifecycle diagnostics are shared, and evaluator state is retry-safe through `state plan record-evaluation`.
-
-## [3.0.0] · 2026-07-12
-
-### Key highlights
-
-- **One runtime lifecycle contract**: OpenCode, Codex, Cursor, and GitHub Copilot now share one authority, one canonical skill at `~/.agents/skills/agentera`, one eight-category diagnosis model, and one ownership-safe operation engine.
-- **Bounded status, detailed diagnosis**: `agentera prime` exposes a compact four-runtime summary while `agentera doctor` reports complete evidence, support-floor gaps, ownership state, and exact host-native actions from the same snapshot.
-- **Previewable, convergent repair**: `agentera upgrade --runtime all|opencode|codex|cursor|copilot --dry-run` is side-effect free; rerunning the selection with `--yes` applies only declared Agentera-owned resources and safely resumes partial work.
-- **Claude retirement and one Cursor identity**: Claude Code active support and package manifests are removed. Cursor Agent CLI is required and Cursor IDE is conditional beneath the single `cursor` runtime identity.
-
-### Added
-
-- Added runtime lifecycle schemas and adapters for skills, plugins, hooks, agents, configuration, enablement, trust, and native actions across all four active runtimes.
-- Added append-only ownership-journal recovery, exact operation dependencies, Linux secure directory-relative publication, and outcomes `applied`, `noop`, `failed`, `blocked_unowned`, `skipped_dependency`, and `action_required`.
-- Added explicit retired cleanup through `agentera upgrade --legacy-cleanup claude --dry-run|--yes`; it can remove only the ledger-owned legacy Agentera skill link and excludes user data.
-- Added typed artifact writes under `agentera state`: progress/decision/audit appends, decision satisfaction updates, plan create/task/status/archive lifecycle operations, schema-derived `explain`, server-assigned identifiers, exact retry replay, dry-run diffs, project locking, strict mapped-path safety, staged 10/40/50 compaction, final-byte validation, and atomic publication.
-- Added `install_track` to `prime --format json` (`app_home.install_track`) and `doctor --format json`. Values are `v2`, `v3`, `source`, and `unknown`.
-
-### Removed
-
-- Removed Claude Code as an active runtime, runtime selector, installation target, plugin marketplace, and default analytics source. Historical transcript parsing remains available only with explicit `--import-source claude` consent and inactive provenance.
-- Removed `cursor-agent` as an independent runtime identity. Its binary and corpus source now map to the single `cursor` identity.
-- Removed `rootStatus` from public JSON output (`prime --format json` and `doctor --format json`). The field remains internal for classifier logic but is no longer exposed in diagnostic output.
-
-### Changed
-
 - Changed lifecycle apply so native installation, update, authentication, enablement, and trust are always user-owned `action_required` steps. Automatic filesystem apply is Linux-only; other platforms retain complete preview and diagnosis.
 - Changed the npm package to include every declared OpenCode, Codex, Cursor, and GitHub Copilot manifest/hook/agent source and to reject package, documentation, runtime-ID, or version parity drift during validation.
 - Renamed the top-level `bundle` field to `app` in `prime --format json`. Nested names are now `activeAppRoot`, `cliApp`, and `legacyAppRoot`; this is a breaking JSON output change.
 - Added v2/v3 coexistence warning to the orientation attention list when a v2 managed app is detected at the app home alongside a v3 CLI runtime. Uses the existing "pick one line" contract message from `references/cli/coexistence-probe.yaml`.
 - Promoted English names to canonical v3 capability IDs (`status`, `vision`, `discuss`, `research`, `plan`, `build`, `optimize`, `audit`, `document`, `profile`, `design`, `orchestrate`) and retired Swedish `-era` IDs from the active command surface.
 - Removed top-level `hej`, `describe`, `gate`, and routine-state names from the v3 command tree; use `prime`, `schema`, `check`, and `state` namespaces.
+- Replaced the Python managed-app entrypoint on the development channel with the npm self-contained model; `upgrade --channel development` rewires runtime hooks to `npx -y agentera@next` and runs irreversible v2→v3 cleanup when previewed with `--dry-run` and applied with `--yes`.
+- Enforced the single-name protocol boundary: capability instructions live in `packages/cli/src/capabilities/<name>/instructions.ts` (`instruction_module` in the capability contract), startup JSON is `agentera prime --context <name> --format json`, and storage paths use canonical `.agentera/*.yaml` names instead of legacy Markdown artifact filenames in protocol surfaces.
+- Split `capabilityContext`, `startupAnalysis`, `extractCorpus`, `doctor`, `state`, `codex`, `compaction`, and `validateArtifact` into focused submodules so every `packages/cli/src/**/*.ts` file stays at or under 1000 lines.
+- Grouped CLI help and dispatch under audience namespaces (`state`, `report`, `check`) with `prime` as the default orientation entry; `agentera --help` lists only canonical top-level commands while legacy invocations still delegate with stderr deprecation.
 
 ### Fixed
 
+- Fixed the npm shim to diagnose a configured `AGENTERA_HOME` whose managed script is missing or not runnable before continuing through repository and `uvx` fallback resolution.
+- Fixed `agentera check lint` to report findings without failing by default, fail the same findings with `--strict`, and keep invalid input, authority failures, and validated plan publication nonzero.
+- Fixed schema/runtime alias parity: introspection now advertises exactly the six Phase 3 transitional aliases, Build startup and compaction guidance use canonical `agentera check compact`, check/fix output reports `command: "check compact"`, top-level `compact` remains behavior-compatible, and removed names such as `gate` return structured runnable corrections.
+- Fixed archive-only plan catalogs to expose executable list/get recovery, plan-task cursors to validate explicit plan selectors before lookup, and legacy plan text to omit oversized whole rows instead of truncating Unicode scalars.
+- Fixed bounded plan and experiment retrieval to declare legacy projection omissions with public recovery, enforce plan-task YAML byte limits, reject wrong-family and wrong-objective cursors, and fail closed when plan snapshot diagnostics change.
+- Fixed experiment snapshot cursors to survive append-driven 10/40/50 full-to-summary and summary-to-drop rollover when immutable archives prove original identities unchanged, while still rejecting mutations, deletion, archive damage, and unsubstantiated rollover.
+- Fixed `packages/cli/src/state/archivePublication.ts` to fsync each newly created archive directory entry and reconstruct missing experiment archives from exact full projection replays before success.
+- Fixed generic `agentera state experiments explain --format json` discovery to return the valid `publish` operation instead of rejecting an invalid `append` default.
+- Fixed plan-list entries to declare full-detail availability, enforced the 32 KiB serialized output budget for YAML as well as JSON, and synchronized retrieval implementation status across authority, schema introspection, help, and runtime behavior.
+- Fixed `agentera state plan --help` to match task-get authority and runtime behavior: `--plan` is optional and omission defaults retrieval to the active plan.
+- Fixed TODO compaction to require one `## ✓ Resolved` section, retain the
+  newest 10 full-detail and 40 summarized rows before dropping older rows, and
+  clarify that Git recovery applies only to previously committed content.
+- Fixed prime dashboard, sparse, and capability startup output to truncate strings at Unicode code-point boundaries instead of splitting emoji surrogate pairs.
+- Fixed full-plan task guidance and startup metadata to use lifecycle and coherence boundaries instead of a fixed task-count ceiling; typed publication still validates task identity and dependencies.
+- Fixed `agentera state plan create` and `archive` publication to use stable archive identities across retries, validate final serialized bytes with strict prose lint and schema checks, and retain canonical `overall acceptance` vocabulary.
+- Fixed `agentera state plan`: `set-status --task` accepts task values, and `set-plan-status` accepts `open` or `complete`.
+- Fixed plan discovery to derive activity from canonical paths: archive-only history returns a complete empty current-plan state with a bounded history catalog, cannot trigger a Build completion sweep, and no longer degrades into omitted current tasks without a retrieval route; next work is dependency-ready, lifecycle diagnostics are shared, and evaluator state is retry-safe through `state plan record-evaluation`.
 - Fixed lifecycle retries after partial writes, stale locks, ownership-journal publication interruption, malformed/forked history, and dependency failures so completed operations remain safe and pending operations converge without adopting user-owned files.
 - Fixed state writes rejecting legacy scalar health archives, accepting missing or circular plan dependencies, following symlinked lock/archive directories outside the project, and advertising invalid plan mutation examples.
 - Fixed `cliDistributionMajor` silently defaulting to v2 when `registry.json` version-authority was missing or corrupted. The function now reads `package.json#version` as a secondary signal before falling back to 2, and emits a one-time stderr diagnostic when both primary and secondary signals miss — preventing v3 installs with corrupted registry assets from being misclassified as v2 in runtime inference, cross-major boundary detection, channel resolution, and next-major doctor rendering.
 - Fixed `pnpm install` failing on hosts with a system-wide `vips` (Arch, Fedora, Nix with vips overlay, Homebrew on macOS): sharp's `useGlobalLibvips()` check returned true and the install hook fell through to a `node-gyp` source build that needed `node-addon-api`. Migrated the deprecated `pnpm.onlyBuiltDependencies` and `pnpm.overrides` from `package.json#pnpm` to `pnpm-workspace.yaml` (pnpm 10+ ignores the legacy key) and added a `preinstall` hook that re-execs `pnpm install` with `SHARP_IGNORE_GLOBAL_LIBVIPS=true` so sharp's prebuilt binary is always used.
 - Fixed `pnpm install` emitting `Failed to replace env in config: ${NPM_TOKEN}` on every invocation: scoped the npm auth config from repo root to `packages/cli/.npmrc`. Both publishing packages (the `agentera` 3.0 CLI on `@next` and the `agentera` 0.0.x shim on `@latest`) find the auth via npm's parent-directory walk; `pnpm install` from the repo root no longer reads any auth-bearing `.npmrc`.
-
-### Migration notes
-
-- **npm shim 0.0.2 v3 deprecation hint**: the `agentera@0.0.2` npm shim on `@latest` now emits a three-line stderr banner on every invocation pointing users to `npx -y agentera@next prime` for the v3 TypeScript CLI. `--help` lists the `@next` command; the install-help failure path leads with the `@next` pointer; `--version` stays banner-free so `--json` and pipe consumers stay clean. Set `AGENTERA_NO_V3_HINT=1` to suppress for CI and scripted invocations.
-- **npm shim 0.0.0 cwd preservation**: the `agentera@0.0.0` npm shim on `@latest` now forwards the user's working directory to the installed Python CLI for the `app-home` backend, so `npx -y agentera` state and prime reads resolve the project's `.agentera/` artifacts instead of the app home. The shim remains the 2.x transitional entry point; the native TypeScript CLI is `npx -y agentera@next` until the 3.0 cutover graduates `@latest`.
-
-## [3.0.0-dev.0] · 2026-06-04
-
-### Key highlights
-
-- **npm TypeScript CLI**: `npx -y agentera@next` ships a self-contained package (`dist/bin/agentera.js` plus `bundle/` with skills, schemas, and `registry.json`) with no repo checkout, no Python, and no `AGENTERA_HOME` for normal use; `agentera prime`, `agentera state <name>`, `agentera check validate`, `agentera doctor`, and `agentera upgrade` run from the published tarball.
-- **Audience-namespace command tree**: top-level `--help` groups `prime`, `schema`, `state`, eleven capability routing commands, `upgrade`, `doctor`, `report`, and `check`; legacy top-level names (`hej`, `describe`, `plan`, `validate`, …) still forward with a one-line stderr alias to the canonical command.
-- **v2→v3 migration preflight**: v3 `upgrade --channel development` reads `{app_home}/v3-handoff.json` when present (written by the stable v2 CLI) to classify preserved user state in under 100ms, with a full app-home scan fallback when the manifest is missing or invalid.
-- **Coexistence doctor warnings**: `agentera doctor` prepends a `Coexistence` section when a v2 managed app home sits beside a v3 npm install (and the stable v2 CLI emits the symmetric warning when npm `@next` is present); both use `references/cli/coexistence-probe.yaml` with three resolution choices.
-- **Bun single-binary build**: `bash scripts/single-binary.sh` compiles `packages/cli/dist/bin/agentera.js` into `packages/cli/dist/bin/agentera-single-binary` after `pnpm -C packages/cli build` and `bundle:data`, for drop-in executables without a Node install.
-
-### Added
-
-- Added the native TypeScript CLI on the npm development channel (`npx -y agentera@next`), bundling app data at pack time via `prepack` (`tsc` + `scripts/copy-bundle.mjs`) and the `.agentera-npx-bundle.json` sentinel checked by `packages/cli/test/packaging/prepack.test.ts`.
-- Added `bash scripts/single-binary.sh` and the packaging contract in `docs/packaging/v3-packaging.md` for Bun `--compile` single-binary delivery alongside the npm tarball.
-- Added `packages/cli/src/migrate/v2HandoffManifest.ts` to read `v3-handoff.json` during v2→v3 cleanup preflight per `references/cli/v3-handoff-manifest.schema.yaml`.
-- Added v3 `agentera doctor` coexistence detection in `packages/cli/src/upgrade/coexistenceProbe.ts` with regression coverage in `packages/cli/test/cli/coexistenceProbe.test.ts`.
-- Added maintainer gates under `agentera check` (`validate`, `verify`, `lint`, `compact`) with JSON parity oracles for validate families, verify eval targets, invalid-input envelopes, and `source_contract` emissions on `prime`, `prime --context`, `state plan`, and `state query --list-artifacts`.
-- Added per-channel `next_major` metadata in `references/cli/update-channels.yaml` and a six-line successor advisory at the head of `agentera doctor` on the stable channel, sourced from `packages/cli/src/upgrade/nextMajorDoctor.ts`.
-- Added a 1000-line monolith lint regression in `packages/cli/test/lint/monolithLint.test.ts` after splitting the eight largest `packages/cli/src/**/*.ts` modules by responsibility.
-
-### Changed
-
-- Replaced the Python managed-app entrypoint on the development channel with the npm self-contained model; `upgrade --channel development` rewires runtime hooks to `npx -y agentera@next` and runs irreversible v2→v3 cleanup when previewed with `--dry-run` and applied with `--yes`.
-- Enforced the single-name protocol boundary: capability instructions live in `packages/cli/src/capabilities/<name>/instructions.ts` (`instruction_module` in the capability contract), startup JSON is `agentera prime --context <name> --format json`, and storage paths use canonical `.agentera/*.yaml` names instead of legacy Markdown artifact filenames in protocol surfaces.
-- Split `capabilityContext`, `startupAnalysis`, `extractCorpus`, `doctor`, `state`, `codex`, `compaction`, and `validateArtifact` into focused submodules so every `packages/cli/src/**/*.ts` file stays at or under 1000 lines.
-- Grouped CLI help and dispatch under audience namespaces (`state`, `report`, `check`) with `prime` as the default orientation entry; `agentera --help` lists only canonical top-level commands while legacy invocations still delegate with stderr deprecation.
-
-### Fixed
-
 - Fixed v3 `agentera upgrade` skipping in-tree `.cursor/agents/*.md` copy-agent steps when the project already uses v3 capability instruction modules, so managed Cursor agents keep `Run agentera prime --context <name> --format json` instead of regressing to `Read …/capabilities/<name>/instructions.md`.
 - Fixed `agentera prime --context orkestrera --format json` returning empty `dependency_ready_tasks` when plan `depends_on` references were complete but normalized differently from task numbers.
 - Fixed v2→v3 migration tests leaking OpenCode skill symlinks into the developer's real home by requiring `MigrationContext.env` in `planRuntimeMigrationItems` and adding `noHomeLeak.test.ts`.
@@ -116,10 +110,18 @@
 
 ### Removed
 
+- Removed Claude Code as an active runtime, runtime selector, installation target, plugin marketplace, and default analytics source. Historical transcript parsing remains available only with explicit `--import-source claude` consent and inactive provenance.
+- Removed `cursor-agent` as an independent runtime identity. Its binary and corpus source now map to the single `cursor` identity.
+- Removed `rootStatus` from public JSON output (`prime --format json` and `doctor --format json`). The field remains internal for classifier logic but is no longer exposed in diagnostic output.
 - Removed v1 legacy migration paths, standalone skill bundle wording, and obsolete v1 fixture assumptions from the 3.x tree after the 3.0 boundary (`packages/cli/test/cli/v1LegacyCruft.test.ts` guards against reintroduction).
-- Removed `hej`, `describe`, and `gate` from top-level `agentera --help`; use `agentera prime`, `agentera schema`, and `agentera check compact` instead (legacy names still run with a stderr alias).
+- Removed `hej`, `describe`, and `gate` from top-level `agentera --help` and dispatch; unsupported structured requests return runnable corrections to `agentera prime`, `agentera schema`, and `agentera check compact`.
 - Removed the requirement for `uv` or a git checkout on the development channel; `uvx --from git+https://github.com/jgabor/agentera@main agentera` remains the stable 2.x Python line only.
 
+### Migration notes
+
+- **npm shim 0.0.2 v3 deprecation hint**: the `agentera@0.0.2` npm shim on `@latest` now emits a three-line stderr banner on every invocation pointing users to `npx -y agentera@next prime` for the v3 TypeScript CLI. `--help` lists the `@next` command; the install-help failure path leads with the `@next` pointer; `--version` stays banner-free so `--json` and pipe consumers stay clean. Set `AGENTERA_NO_V3_HINT=1` to suppress for CI and scripted invocations.
+- **npm shim 0.0.0 cwd preservation**: the `agentera@0.0.0` npm shim on `@latest` now forwards the user's working directory to the installed Python CLI for the `app-home` backend, so `npx -y agentera` state and prime reads resolve the project's `.agentera/` artifacts instead of the app home. The shim remains the 2.x transitional entry point; the native TypeScript CLI is `npx -y agentera@next` until the 3.0 cutover graduates `@latest`.
+
 ## [2.7.7] · 2026-06-03
 
 ### Changed
@@ -460,7 +462,7 @@
 - 2.3.4 patch release readiness is recorded locally without publication, installed app refresh, tag, or remote push; version-bearing package, plugin, registry, lockfile, OpenCode marker, and skill frontmatter surfaces are aligned.
 - PLAN artifact fallback closeout is synchronized without a selected release target:
   focused CLI/contract tests passed with 88 selected tests, `agentera plan
-  --format json --fields source_contract` reports
+--format json --fields source_contract` reports
   `complete_for_plan_artifact=true` and `raw_artifact_reads_required=false`, the
   mixed-window benchmark recorded `245` estimated redundant tokens saved versus
   the previous row, and a clean post-fix probe produced zero raw/redundant

package/bundle/README.md

@@ -93,6 +93,23 @@ agentera state query --list-artifacts
 agentera state progress explain --verb append --format json
 ```
 
+Bounded collection retrieval uses stable identities, opaque snapshot cursors,
+explicit omissions, and exact detail commands. The runtime contract is
+[`references/artifacts/state-storage-authority.yaml`](./references/artifacts/state-storage-authority.yaml).
+
+```bash
+agentera state plan list --format json
+agentera state plan get --plan PLAN_ID --format json
+agentera state plan tasks list --limit 20 --format json
+agentera state experiments list --objective OBJECTIVE_ID --format json
+agentera state experiments get --objective OBJECTIVE_ID --number N --format json
+```
+
+Plan list/get spans the active plan and immutable plan archives. Plan task
+list/get is active-plan-only. Experiment retrieval merges its bounded projection
+with objective-owned immutable archives and reports detail as full, summary-only,
+or unavailable.
+
 The standard project state is:
 
 - `.agentera/vision.yaml` — product direction

package/bundle/UPGRADE.md

@@ -191,6 +191,44 @@ not select an Agentera runtime identity or limit lifecycle diagnosis. Use
 combined with `--only`. Development-channel dry-runs without `--runtime` still
 observe all active lifecycle runtimes.
 
+## Legacy state and optional Git enrichment
+
+Project-state migration and historical Git enrichment are separate local
+operations. Migration does not read Git, contact a remote, or remove its source
+files. Its default inventory and `--dry-run` are read-only; apply requires the
+exact artifact and entry selectors with `--apply --force`:
+
+```bash
+npx -y agentera@next state migrate --project "$PWD" --artifact progress --number N --dry-run --format json
+npx -y agentera@next state migrate --project "$PWD" --artifact progress --number N --apply --force --format json
+```
+
+Git enrichment is optional. A preview is useful but not required before direct
+apply. Inventory and preview may use the current working directory when
+`--project` is omitted; direct apply must name `--project PATH` explicitly.
+Apply revalidates the selected project, current `HEAD`, allowed-ref
+reachability, candidate provenance and content, and the immutable archive target
+immediately before publication:
+
+```bash
+npx -y agentera@next state backfill --project "$PWD" --artifact progress --number N --dry-run --format json
+npx -y agentera@next state backfill --project "$PWD" --artifact progress --number N --apply --force --format json
+```
+
+The backfill contract is bounded to 100 result rows, 500 history units, and
+16 MiB of Git output. It inspects only `HEAD`, local heads, and tags; remote and
+custom refs and operations are excluded. Returned provenance keeps the commit,
+path, blob ID, stable entry ID, content hash, and reachability so a result can
+be traced without adding commit fields to the archive record.
+
+`complete`, `degraded`, `blocked`, and `unavailable` are explicit outcomes.
+Changed, shallow, rewritten, ambiguous, corrupt, missing, bounded, unavailable,
+and immutable-conflict results are reported rather than guessed. Resolve the
+reported local condition and retry the same selectors; refused operations leave
+active projections and existing archives unchanged, and identical publication
+converges as a replay. The full command, limits, failure, recovery, and
+traceability contract is `references/artifacts/state-storage-authority.yaml`.
+
 ## Verification and recovery
 
 After apply:

package/bundle/references/adapters/package-surface-characterization.md

@@ -23,6 +23,13 @@ development pre-release suffix, but its `X.Y.Z` core must match the suite:
 `.opencode/package.json` is a runtime package manifest but intentionally has no
 suite-version field.
 
+`packages/cli/package.json#agentera.gitRef` identifies the last substantive
+package-source commit, not a later verification-only commit. Repository-local
+release validation requires that commit to exist and compares the package
+contract, compiled-source inputs, scripts, and bundled-data inputs against it.
+Only the package version and `gitRef`, project state, and the release validator
+implementing this check are excluded, avoiding a circular source reference.
+
 ## Active runtime manifests
 
 The runtime manifest set covers exactly OpenCode, Codex, Cursor, and GitHub