Agentic Praxis Grimoire

The marker above is the mechanical current-state authority. Bounded contradiction diagnostics cover only their frozen vocabulary; arbitrary prose still requires human review.

Agentic Praxis Grimoire (APG) is a modular engineering operating model for coding agents. It curates practices that improve planning, implementation, testing, debugging, review, delivery, and coordination without turning one source methodology into a universal workflow.

Public APG v0.4.0 contains twenty-eight skills, projection and reporting components, source-specific release and user-lifecycle validation, governance, and licensing terms. Fourteen catalog rows are stable and fourteen remain provisional; release inclusion does not change maturity. v0.4.0 appends one intentionally squashed release commit and an annotated tag to the preserved v0.1.0 through v0.3.0 history. Its canonical public checkout supplies the maintainer's separately managed user-global Codex integration. Private development history remains distinct.

APG49 independently validates the APG48 matryer/is v1.4.1 candidate but does not retain it: fresh review after the single correction pass finds new material source, trigger, structural, and scenario-continuity defects. ADR 0030 is Rejected, ADR 0026 remains the controlling two-component no-stack Go architecture, and development remains 28/28/28. See the APG49 evaluation.

APG59 independently validates and amends the APG58 CSS pilot, but corrected- state review finds material executable-contract and removal defects after the single correction pass. ADR 0035 is Rejected, all candidate surfaces are removed, and development remains 28/28/28 with fourteen stable and fourteen provisional rows. Corrected public and active v0.4.0 remain unchanged.

APG60 now freezes the rejected-evidence behavior before any re-authoring: sixty structured CSS response cases, a deterministic reference evaluator, and a generic candidate-removal model exercise every current owner and preserve history. It authors no skill, leaves ADR 0036 unused, and grants no successor authority.

APG0 through APG8 are the closed v0.1 development epic, with each historical terminal outcome preserved, including APG3's blocked result. The epic established the project, bootstrap maturity model, six skills, repository discovery, dogfooding, project-local projection and rollback, and RepoMap managed adoption. The maintainer subsequently decommissioned Superpowers, and a bounded fresh RepoMap smoke passed after decommission. APG9 closes that epic, reconciles current state, and accepts the APG10 through APG14 v0.2 roadmap in ADR 0006.

APG10 accepts that closeout, resolves the experimental guideline source's provenance and reuse boundary, and records its final dispositions in ADR 0007. Frozen scenarios found current assumption, alternative, traceability, and speculative-scope behavior adequate. They demonstrated one narrow gap for locally owned code or test artifacts made unnecessary by the authorized change, which received one independently reviewed implementation-discipline correction. No seventh skill or maturity change was introduced.

APG11 accepts ADR 0008, establishes the skill authoring and maintenance guide, adds a dependency-free read-only mechanical skill-library checker, and closes the former candidate-theme queue through the legacy roadmap ledger. It changes no skill leaf or maturity state. APG11A's accepted lexical correction then closes required-key shadowing and fenced-code false negatives without changing that architecture.

APG12 accepts ADR 0009, adds an exact non-private public surface policy, and supplies separate dependency-free commands for reproducible local public candidates and state-owned user-scope skill links. Its executable regression detects the public v0.1.0 omitted-wrapper class. Disposable candidate and user lifecycle dogfood leave the public repository and active integration unchanged. APG12A subsequently corrects public-lineage and read-only validation defects without reopening that architecture.

APG13 accepts ADR 0010 after individual historical inventories, frozen current applications, complete regression, and fresh non-author reviews. All six current catalog entries are stable; every canonical leaf remains byte-identical to the APG12A baseline. APG14 corrects two APG9 evidence labels without changing those dispositions, publishes the exact non-private v0.2.0 projection, and fast-forwards the active public-backed source without changing its integration ownership shape. Public v0.1.0 remains historical and unchanged. Stability and publication do not establish clean comparative superiority, production warranty, universal applicability, or automatic invocation. Superpowers remains retired. The maintainer subsequently completed the requested full Codex restart and fresh-session discovery smoke and reported that it passed.

APG15's capability-oriented, synthesis-first v0.3 architecture is accepted in ADR 0011. APG16 adds one provisional public workflow router for ambiguous selection, routing audit, and capability-health diagnosis. It routes to one smallest sufficient stable process leaf or none; it is not a session bootstrap, mandatory chain, action authority, or procedure owner. The maintainer's later fresh-session observation closes its duplicate-name discovery request.

APG17 accepts ADR 0013 and adds one provisional synthesizing-repository-guidance leaf. It classifies mixed guidance into bounded owners, rights, privacy, migration, and rollback dispositions before any rewrite. It does not author the resulting artifact, mirror private sources, migrate root guidance, remove private skills, or grant implementation authority. APG18 accepts ADR 0012, adds the normative language-profile contract and one provisional python-language-profile, and preserves the six-skill public v0.2 lifecycle. APG19 accepts separate shell-language and test-harness ownership, adds provisional Bash, Bats, and Zsh profiles, and defers ZUnit on current source/runtime evidence.

APG19A accepts that substantive result and adopts semantic phase identity, independent ADR and exit sequences, semantic durable references, and precommit record finalization in ADR 0015. Its focused audit corrects Bats test counting for the runner-supported comment function form without changing thresholds, maturity, catalog shape, the six-skill v0.2 lifecycle, or the APG23/APG24 smoke deferral.

APG20 truthfully defers independent Go and Ruby profile candidates after fresh review identifies material semantic, measurement, source-license, and dogfood defects beyond its correction allowance. APG20A accepts that defect ledger, corrects the report-lock race exposed by APG20's retry, and retains corrected Go and Ruby profiles as provisional development skills. APG21 accepts separate Nix, PostgreSQL, and SQLite ownership under ADR 0016 and retains PostgreSQL and SQLite as provisional profiles. Nix is deferred-material-defect after its corrected candidate exposes a second behavior-bearing contradiction. APG21A corrects that retained defect ledger, retains Nix provisionally, and applies one bounded PostgreSQL false-escalation correction. Private development now contains nineteen skills and projections: six stable process skills and thirteen provisional v0.3 skills before APG23. The router map contains eighteen routable non-router entries. No generic SQL owner, root or private cutover, or v0.3 release is implemented.

APG22 records 35/35 matched read-only router, retained-profile, and guidance- synthesis dogfood cases across APG, RepoMap, and public-safe synthetic boundaries. It finds no behavior-bearing APG defect and proposes no APG root reduction. The guidance-migration proposal keeps target-specific and private guidance with its current owner and requires distribution, shadow, discovery, override, rollback, and target acceptance before any later cutover. The v0.3 release-scope ledger records APG22A's retained approved-roadmap manager-assignment owner and APG22B's exact ZUnit v0.8.2 with Zsh 5.9.2 profile. The unsupported 5.3.1 pair and every unverified range remain excluded. No cutover, decommission, public release, active-integration change, or application smoke occurs.

APG22C subsequently corrects the ZUnit harness's selected user-startup evidence. Matched positive, negative, and in-runner controls retain the exact 5.9.2 support boundary and the 5.3.1 unsupported result without changing a skill, maturity row, catalog shape, public v0.2, or active integration.

APG23 completes the required fresh-session discovery and explicit-use smoke, records independent maturity and release-inclusion decisions for all thirteen v0.3 skills, and accepts all thirteen for v0.3. Eight rows are promoted to stable and five remain provisional, producing fourteen stable and five provisional development rows. Public v0.2, schemas, managed defaults, and the active integration remain unchanged. APG24 is the only remaining v0.3 phase and separately owns candidate construction and publication.

APG24 accepts ADR 0019, publishes all nineteen skills as v0.3.0, expands new project defaults and source-specific user transitions while retaining schema version 1, and fast-forwards the active public-backed source without changing its aggregate link ownership. Public v0.2.0 remains unchanged. The personal same-name router remains installed for a source-qualified fresh-session shadow; no root/private cutover, router decommission, or successor phase is authorized.

APG24A records the maintainer-reported successful public-v0.3 fresh-session shadow and the later personal-router decommission performed under separate human authority. Focused verification confirms unchanged public and active v0.3.0 state, nineteen aggregate skills, preserved personal transition targets, and an exact private restoration source. v0.3 is terminal; APG24A changes only documentation and publication-excluded evidence.

APG25 accepts structured project defaults, testing and coverage policy, a Python-first reporting architecture, ChatGPT-manager topology, and the v0.4 roadmap through ADRs 0020-0022. One bounded correction makes approved-roadmap assignments consume repository defaults and state deviations rather than repeat ordinary procedure. No skill is added, tool is converted, test is migrated, ChatGPT leaf is moved, personal skill is changed, or release is begun.

APG26 adds provisional pytest-test-profile and converting-bash-scripts-to-python private-development capabilities after current-source calibration, sixty frozen scenario families, failing-first focused tests, read-only report-tool dogfood, and fresh review. Development is now 21 canonical skills, 21 catalog rows, 21 relative projections, 14 stable rows, 7 provisional rows, and 20 routable non-router capabilities. Public and active v0.3.0 remain unchanged at nineteen skills; no report tool is converted, no git-diff-report is implemented, and no test is migrated.

APG26A records that APG26's formal-phase commit contains only its subject and corrects forward without rewriting APG26. The dependency-free Python apg-check-phase-commit-message command rejects that regression and validates the required phase subject plus ordered Scope, Result, Verification, and Not run sections before and after new APG formal-phase commits. No skill, maturity, release, active integration, report executable, or report format changes in APG26A.

APG27 accepts the design in ADR 0023 and produces a partial, uncommitted standard-library libexec/agent_report candidate. The Python git-show-report preserves version-2 bytes on the characterized POSIX platform, git-diff-report adds deterministic drift-checked uncommitted state evidence through a private index, and append-operational-report enforces exact same-report Git show/diff association. The phase stops partial before commit because the development release checker does not retain immutable v0.3.0 policy as a historical surface and the Python profile retains a Red branch signal. GitPython is not selected, the pytest migration remains deferred, and the published and active v0.3.0 objects remain unchanged.

APG27A preserves that partial result, freezes immutable v0.3.0 policy independently from the current development inventory, decomposes the Red path- safety owner without changing diagnostics, and adopts the corrected report core. Focused parity, safety, association, historical-release, and independent review gates pass. APG27A does not begin the pytest migration or change public or active v0.3.0.

APG28A subsequently adopts the corrected pytest migration after preserving APG28's Partial result, and APG29 aligns the affected process owners with the adopted structured-project and test defaults.

APG30 implements ADR 0022's first actor-qualified namespace. The new provisional chatgpt-manager-workflow subrouter and the unchanged composing-approved-roadmap-assignments leaf are canonical under skills/chatgpt/, while both retain flat .agents/skills/<name> discovery links. Development is 22/22/22 with fourteen stable and eight provisional rows. The general map has nineteen ordinary non-ChatGPT leaves plus the subrouter; the ChatGPT-local map has the one manager leaf. Public and active v0.3.0 remain immutable at 19/19/19. Application discovery and every personal skill transition remain behind APG31's mandatory restart gate.

APG31 passes that fresh-session topology gate and independently shadows three personal hygiene capabilities against current APG and repository owners. The personal docs-only capability is decommissioned after replacement, restoration, and non-author review. Git-history scope reduction and RepoMap phase-hygiene decommissioning remain deferred because their current private routing boundaries cannot be corrected within APG31 authority. Development remains 22/22/22, router maps remain exact, and public and active v0.3.0 remain unchanged. A fresh-session post-transition discovery smoke is still required.

APG31A records that later smoke as passed while preserving APG31's historical Partial result. It completes the two deferred dispositions: git-history-hygiene is scope-reduced after generalized behavior returns to APG and repository owners, and repomap-phase-hygiene is decommissioned after generalized behavior returns to current RepoMap owners. APG development and public/active v0.3.0 remain unchanged.

APG32 retains the provisional minitest-test-profile after current Minitest, Ruby, and extracted mock calibration; thirty-six frozen trigger, non-trigger, semantic, structural, and stop families; a failing-first mirrored contract; and fresh non-author review. One bounded correction makes fixture alternatives and Minitest-specific boundary effects explicit while preserving existing general, Ruby, and repository owners. Development becomes 23/23/23 with fourteen stable and nine provisional rows. The general map has twenty-one edges; the ChatGPT-local map remains one edge. Public and active v0.3.0 remain immutable at 19/19/19. No dependency, framework selection, readiness, smoke, release, publication, or successor phase is included.

APG33 retains the provisional dockerfile-profile after current Dockerfile frontend, BuildKit, Docker documentation, and OCI image-configuration calibration; forty frozen trigger, non-trigger, semantic, structural, and stop families; a failing-first mirrored contract; and fresh non-author review. The profile owns Dockerfile-specific parser, stage, instruction, context, copy, mount, cache, user, metadata, and platform judgment while preserving image, dependency, runtime, release, and live-operation authority. Development becomes 24/24/24 with fourteen stable and ten provisional rows. The general map has twenty-two edges; the ChatGPT-local map remains one edge. Public and active v0.3.0 remain immutable at 19/19/19. No Docker build, container, daemon, registry, readiness, smoke, release, or publication action is included.

APG34 retains the provisional vagrantfile-profile after current Vagrant source and documentation, configuration-load, Ruby-compatibility, box, provider, plugin, network, synced-folder, provisioner, trigger, and state calibration; forty frozen trigger, non-trigger, semantic, structural, and stop families; a failing-first mirrored contract; one bounded source-semantics and machine-measurement correction; and fresh non-author review. The profile owns Vagrantfile-specific configuration judgment while preserving provider, box, plugin, host, network, filesystem, command, lifecycle, release, and live-operation authority. Development becomes 25/25/25 with fourteen stable and eleven provisional rows. The general map has twenty-three edges; the ChatGPT-local map remains one edge. Public and active v0.3.0 remain immutable at 19/19/19. No Vagrantfile evaluation, box or plugin mutation, provider contact, machine lifecycle, readiness, smoke, release, or publication action is included.

APG38 retains provisional go-test-profile and go-cmp-test-profile after independent current-source review, 66 public-safe scenario families, categorical corpus calibration, isolated Go 1.25.10 compatibility probes, and fresh corrected-state review. ADR 0026 accepts two directly triggerable Go component owners without a stack owner. matryer-is-test-profile and nix-test-profile are deferred and absent after new post-correction attribution/false-escalation and FreeBSD sandbox-default defects. Development becomes 27/27/27 with fourteen stable and thirteen provisional rows. The general map has twenty-five edges; the ChatGPT-local map remains one edge. Public and active v0.3.0 remain immutable at 19/19/19. No Nix execution, target-repository test, readiness, smoke, release, publication, deployment, or successor phase is included.

APG40 independently integrates the APG39 replacements. nix-test-profile passes exact Nix 2.35.1 and pinned Nixpkgs/NixOS 26.05 source review, forty corrected public-safe scenarios, source-only structural calibration, one coherent correction cycle, and fresh non-author review, then begins provisional. matryer-is-test-profile is deferred-material-defect after corrected-state review finds its equality mechanism still inaccurate. ADR 0027 is Rejected; ADR 0026 remains Accepted and controlling; no go-testing-stack exists. Development becomes 28/28/28 with fourteen stable and fourteen provisional rows, twenty-six general-map entries, one ChatGPT-local entry, and twenty-seven checked route edges. Public and active v0.3.0 remain immutable at 19/19/19.

APG41 closes the v0.4 development surface with all fourteen provisional rows retained and unpromoted. Cross-profile dogfood preserves direct owner selection and no mandatory chain; one bounded correction makes the Minitest, Dockerfile, and Vagrantfile removal descriptions candidate-independent. Two disposable exact v0.4.0 candidates and isolated lifecycle smoke pass on the current host. The terminal result is ready-for-publication-with-provisional-limitations; it publishes and deploys nothing, and it authorizes no successor phase. Public and active v0.3.0 remain 19/19/19.

Authority

The human maintainer retains ultimate project, roadmap, publication, license, and destructive-action authority. ChatGPT manages planning and review only within a human-authorized task, phase, or preapproved roadmap envelope. Top-level Codex executes bounded ChatGPT assignments and manages internal Codex workers. Evidence and recommendations do not expand any actor's authority. The manager-worker protocol defines the complete chain and stop boundaries.

Intended audience

APG is for maintainers who design agent workflows, agents that implement or review those workflows, and contributors evaluating whether a practice improves correctness, safety, maintainability, or delivery outcomes at an acceptable cost.

Project structure

• AGENTS.md contains the small set of repository-wide rules that should apply to nearly all work.
• docs/project-model.md defines artifact ownership, evidence domains, and the candidate-to-adoption lifecycle.
• docs/skill-authoring-and-maintenance.md owns the APG skill-specific authoring, correction, support, maturity, deprecation, and removal procedure.
• docs/provenance.md defines public and publication-excluded provenance responsibilities.
• docs/manager-worker-protocol.md defines external authority, top-level management, internal delegation, evidence, and final reporting.
• docs/structured-project-phase-defaults.md defines formal and non-phase commit, status, ADR, docs-only, scoped-test, and manager-assignment compression defaults.
• docs/testing-and-coverage-policy.md defines scoped tests, coverage remediation, mock boundaries, and the adopted APG pytest architecture.
• docs/agent-reporting-architecture.md defines the Python-first Git-show, Git-diff, and operational-report ownership and format boundaries.
• docs/chatgpt-manager-skill-topology.md defines implemented nested ChatGPT canonical owners, flat discovery, subrouting, and personal transition gates.
• docs/adr/ records durable architecture decisions under an independent four-digit sequence.
• docs/status/ records truthful phase exits under an independent five-digit sequence.
• skills/ contains and indexes thirty-three canonical skills: fourteen stable rows and nineteen provisional manager-assignment, language, database, test-profile, or conversion rows.
• .agents/skills/ is the checked-in Codex repository discovery projection; its thirty-three relative symbolic links contain no independent skill content.
• docs/bootstrap-v0.1.md defines maturity, provisional evidence, rollback, dogfooding, and decommission gates.
• docs/superpowers-transition.md maps materially relevant Superpowers workflows to APG, native Codex, project policy, deferral, or rejection.
• docs/project-skill-projection.md documents opt-in cross-repository installation, adoption, verification, and rollback.
• docs/public-release-process.md documents exact projection, deterministic candidate construction, validation, and the v0.2.0 and v0.3.0 publication records.
• docs/user-scoped-skill-integration.md documents direct public-sourced user links, state, lifecycle, restart, and migration boundaries.
• docs/superpowers-decommission-runbook.md preserves the human-owned decommission and rollback sequence after the completed operation without authorizing restoration.
• docs/evaluations/ records the public-safe APG4 scenario and review summary; the APG3 blocked record remains preserved separately, and the APG10 evaluation records the experimental-source dispositions.
• docs/roadmap.md records the closed v0.1 and v0.2 sequences, the completed v0.3 sequence, APG25's v0.4 foundation, APG26's initial enabling capabilities, and APG26A's formal-commit correction. The detailed v0.4 roadmap owns the dependency-ordered future slices without allocating phase IDs.
• docs/legacy-roadmap-closure.md gives every former candidate or deferred theme a terminal owner or condition.
• bin/ and libexec/ contain deterministic reporting and change-size tools, the dependency-free project-local, user-scoped, and flat shared-skill projection commands, the read-only skill-library, record-identity, and formal-phase commit-message checkers, and the local-only public candidate builder/checker with non-executable helpers.
• src/test/unit/python/agentic-praxis-grimoire/ and src/test/int/python/agentic-praxis-grimoire/ mirror production owners for isolated unit tests and real-boundary integration tests. Test filenames use .unit.test.py or .int.test.py suffixes.

Testing

The repository separates tests by level and implementation language:

src/test/unit/<language>/
src/test/int/<language>/

APG28A adopts the corrected pytest migration. The repository interfaces are:

bin/apg-test unit
bin/apg-test integration
bin/apg-test unit-integration

Each command defaults to eight xdist workers. The strict source and mirror inventory is recorded in testing/apg-test-inventory.json; exact integer statement and branch counts enforce 80/80 component and 85/85 union gates. Run-scoped manifests account for xdist workers, collection, terminal results, and required Python-child coverage. Both report-tool Bats files remain.

Validate the current canonical skill library and checked-in Codex projection without mutation:

bin/apg-check-skill-library [--root <path>] [--format text|json]
bin/apg-check-record-identity [--root <path>] [--format text|json] \
  [--expect-available <phase>] [--expect-allocated <phase>]
bin/apg-check-phase-commit-message --phase <PHASE-ID> \
  (--message-file <path> | --commit <revision>) [--format text|json]
bin/apg-check-change-size staged [--format text|json]
bin/apg-check-change-size commit <commit> [--format text|json]
bin/apg-check-change-size tree [<commit>] [--format text|json]

These commands validate only their adopted mechanical APG subsets. They do not prove semantic quality, imperative mood, authority, privacy, provenance, client discovery, maturity, release completeness, record truth, or stable behavior.

The adopted Python reporting interfaces are:

bin/git-show-report <phase-id> <commit> <status-doc> <result> <final-gate>
bin/git-diff-report <phase-id> <result> <final-gate> [--status-doc <path>]
bin/append-operational-report <phase-id> <absolute-body-path> <result> <final-gate> [options]

The three compatibility report commands and canonical apgr report routes default to one current primary under ~/Documents/agent/outbox/<project>/<phase>/<phase>.<kind>.report.txt. GIT_SHOW_REPORT_ROOT remains an explicit legacy-omnibus override for the compatibility commands, but cannot redirect canonical apgr report routes; historical reports under the earlier default are not migrated. apgr report recover --phase <phase> provides bounded interrupted-transaction recovery; repository-backed writes reject an explicit project identifier that does not match the repository basename.

bin/flatten-skill-symlinks INPUT OUTPUT creates a flat, owner-state-backed symlink projection from nested skill roots. It supports deterministic collision handling, --dry-run, --check, state-owned --replace and --clean, and refuses overlap, symlinked output ancestry, unmanaged collisions, and indirect marker files. APG53 did not mutate a live Claude installation.

bin/install-global-skills AGENT [REPOSITORY ...] projects one complete set of local skill repositories into a Codex or Claude personal skill root. With no repository arguments it uses the installed APG repository. Explicit repositories replace that default; --include-apg adds APG to an explicit set. Codex defaults to $HOME/.agents/skills, Claude defaults to ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/skills, and --skills-root overrides the destination exactly. Duplicate names, unmanaged collisions, and other command owners fail closed. --check, --dry-run, and exact owned --uninstall are supported. This is local symlink projection, not package, plugin, trust, release, or runtime-discovery validation.

bin/apg-check-change-size reads index, commit, or tree Git objects. The versioned policy in testing/apg-change-size-policy.json limits ordinary blobs to 256 KiB, generated-derived evidence blobs to 128 KiB, generated evidence added per change to 512 KiB, and text lines to 64 KiB. Archives are rejected by default and binary assets require exact identity-bound exceptions.

On Windows invoke the entry point through an interpreter, for example python bin/git-diff-report --help. Report replacement fails closed there until native sharing, reparse, and replacement safety is characterized. A Git record must exist before its associated operational append. Standalone operational records are accepted only when the canonical phase report contains no Git show or diff record.

From source to APG practice

A candidate practice moves through a bounded lifecycle:

1. inventory dated source evidence and its ownership, publication, and license status;
2. evaluate the concrete problem, evidence strength, generality, activation risk, maintenance cost, and appropriate destination;
3. propose an APG-native change with provenance and observable validation criteria;
4. validate the change proportionally, including representative non-trigger and failure cases where relevant;
5. adopt, defer, reject, or supersede it with a recorded rationale; and
6. keep the adopted artifact, evaluation evidence, documentation, and provenance consistent.

Copied or adapted expression requires confirmed reuse rights and any required notice. Synthesized or inspired practices still retain useful provenance. Source inclusion, frequency, ownership, or apparent authority is not itself an adoption decision.

The project model owns this general lifecycle. The skill authoring and maintenance guide owns its proportional application to skill changes.

Publication model

The canonical public identity is agentic-praxis-grimoire. Public v0.1.0 was published as a filtered projection with one intentionally squashed commit and historically omitted the documented bin/apg-project-skills wrapper. ADR 0009 replaces manual selection with an exact projection of every tracked path except private/, plus critical-owner checks that detect deletion from source. Public v0.2.0, v0.3.0, and v0.4.0 each append one deterministic squashed release commit and annotated tag while preserving the preceding public release as sole parent. v0.4.0 publishes twenty-eight skill owners with fourteen stable and fourteen provisional rows, then advances the aggregate-owned active source by exact fast-forward. Future publication remains a separately authorized human decision.

License

Agentic Praxis Grimoire is licensed under the GNU Affero General Public License v3.0 or later. See LICENSE.

Commercial licenses are available for proprietary terms, including closed-source embedding, private service deployments, OEM use, support, warranty, indemnity, and custom commercial terms. See COMMERCIAL-LICENSE.md.

Contributions are accepted under the terms in CONTRIBUTING.md and CLA.md.

Where to begin

Agents should read AGENTS.md, then the focused owner for the task. Maintainers evaluating source-derived policy should begin with the project model, provenance policy, and the ADR index. The reviewed but rejected Web and Node candidate boundaries and non-normative growth evidence are summarized in the Web and Node profile-family architecture. Its repaired reproducibility foundation and named insufficient evidence classes are recorded in the APG52 evaluation; that evidence accepts no threshold or profile. A later fresh reconstruction on that foundation produced ADR 0034 with separate owner, authoring, and growth-band decisions; see the APG56 evaluation and the reconstruction architecture document. The APG57 independent review rejects ADR 0034 after material corrected-state defects, leaves every numeric band and all authoring deferred, and changes no skill surface. The APG58 CSS pilot authors one unintegrated css-language-profile candidate under explicit policy-selected structural limits. The APG59 independent validation closes projected/resulting-count and incremental-growth loopholes and narrows authority and semantic boundaries, but corrected-state review finds material verification and removal defects. It rejects ADR 0035 and removes the candidate from current development. The APG60 foundation then freezes sixty candidate-independent response cases and a complete retained/rejected removal-closure model before any future author may write candidate prose. It creates no CSS surface and does not authorize APG61. The current multi-repository installer and its forward transaction hardening are recorded in the APG54 integration and APG55 correction. Delegated work should follow the manager-worker protocol. The completed v0.1 epic, post-release APG-TEST0 foundation, and completed APG10 through APG14 v0.2 sequence are described in the roadmap. No successor roadmap epic begins automatically.

APG60A CSS foundation correction

APG60A corrects the APG60 pre-authoring oracle forward without restoring the rejected CSS candidate. Projection overruns are operation-independent, artifact exceptions retain their real authority source, broken symlink residue is lexical, and later retained integration must pass an actual-tree closure gate. The 60-case register and 300/600/900 policy are unchanged.

See APG60A CSS Contract Foundation Hardening. APG61 remains recommended, not authorized.

APG60B traceability and decision closure

APG60B closes five additional pre-authoring false-pass classes without changing the accepted CSS behavior contract. Future contract maps must agree exactly with all sixty cases and reachable clause anchors; narrative state, Python owner variables, derived release values, and ADR 0036 lifecycle state are checked against their real owners. No CSS candidate or ADR 0036 is created, and no successor is authorized.

APG60C runtime and terminal lifecycle closure

APG60C preserves accepted APG60B and records the owner-source finality claim later narrowed forward by APG60D. APG60C keeps exact marker state separate from bounded prose diagnostics, positively gates the authored Proposed but unintegrated state, requires preserved history for terminal rejection, and rejects symlinked or malformed current survivor owners. The sixty-case behavior contract and 28/28/28 development state remain unchanged; APG61 is recommended only and is not begun or authorized.

APG60D source-binding and phase-history closure

APG60D preserves accepted APG60C while replacing its overstrong runtime/call claim with truthful static source-binding integrity. Exact APG58 through APG60D foundation bundles are mandatory; authored state requires APG61, and retained or rejected state requires APG61 plus APG62. All repository history is exact, direct regular, nonempty UTF-8, and no-follow. The CSS contract and 28/28/28 development state remain unchanged. APG61 uses future exit 00085, is recommended only, and is not begun or authorized.

APG60E repository-path and candidate-surface closure

APG60E preserves accepted APG60D and closes lifecycle authority, authored-owner, and retained-owner ancestor-symlink false passes through one physical-root, descriptor-relative no-follow read contract. The expected projection remains an exact relative symlink whose parent and canonical target provenance are now closed. The CSS behavior contract and 28/28/28 development state remain unchanged. APG61 uses future exit 00086, is recommended only, and is not begun or authorized.

APG60F import, owner, and projection closure

APG60F preserves accepted APG60E while closing repository-local transitive import and cache isolation, making all 52 lifecycle roles and cross-references exact, and revalidating the expected projection and canonical target as one bounded observation. APG58 through APG60F are the foundation. CSS remains absent, ADR 0036 remains unused, and development, public, and active state remain unchanged. APG61 uses exit 00087, remains recommended only, and is not begun or authorized.

APG60G snapshot, role, and derived-set closure

APG60G preserves accepted APG60F while binding dynamic consumers to one pinned repository object, comparing complete exact derived skill sets, freezing the 52 required semantic roles in a code-owned registry, and making authoritative regular reads observe the same final repository entry before and after the read. APG58 through APG60G are now the foundation. CSS remains absent, ADR 0035 remains Rejected, ADR 0036 remains unused, and development, public, and active state remain unchanged. APG61 uses exit 00088 and APG62 uses 00089; APG61 remains recommended only and is not begun or authorized.

APG60H snapshot and full-path binding closure

APG60H preserves the immutable Claude object and completes worker scratch, exact snapshot, full-path presence and absence, final root binding, and projection identity verification. APG58 through APG60H are foundation; future exits are 00089 and 00090. Candidate, public, active, target, and successor state remain unchanged.

APG60I worker temporary-root binding and cleanup closure

APG60I preserves APG60H's adopted-with-exceptions history and corrects its two known worker temporary-storage defects forward. Child creation is bound to an already-opened no-follow root and cleanup owns the lifecycle before creation. APG58 through APG60I are foundation; future exits are 00090 and 00091. CSS, ADR 0036, public, active, target, and successor state remain unchanged.

APG61 CSS language-profile authoring

APG61 authors one fresh css-language-profile candidate from the frozen APG60A contract on the preserved Claude authoring branch — leaf, specification, and sixty-case traceability map — and proposes ADR 0036. The candidate is not integrated: current candidate-state markers remain absent, integrated counts remain 28/28/28, development main remains at exact APG60I, and public, active, and target state are unchanged. APG62 validation at exit 00091 is recommended but separately authorized.

APG62 CSS language-profile validation and rejection

APG62 independently reconstructs the frozen sixty-case oracle and validates the exact delivered APG61 candidate. Seven initial semantic-navigation defects receive the one permitted coherent correction; fresh corrected-state review then finds a new material six-case record-growth-state overreach. ADR 0036 is therefore Rejected and all current candidate surfaces are removed. Development remains 28/28/28 with 14 stable / 14 provisional, current markers remain absent, APG61 history is preserved, and public, active, and target state is unchanged. No successor is authorized.

APG63 Markdown architecture and lean contract

APG63 moves to Markdown after the terminal CSS rejection without retrying css-language-profile. From exact APG62 it reverifies CommonMark 0.31.2, the pinned GFM specification, CC BY-SA 4.0 document rights, and the seven target Markdown documents; defines a narrowed coherent Markdown owner with closed raw-HTML, frontmatter, and MDX/host boundaries; selects qualitative structure-first structural policy (no numeric whole-file bands) under ten pre-frozen purpose controls; freezes a lean thirty-six-scenario candidate-independent contract; and proposes ADR 0037. No skill is authored and nothing is integrated: development remains 28/28/28 with 14 stable / 14 provisional, ADR 0036 stays Rejected, and public, active, and target state are unchanged. Codex peer review (recommended APG64) terminally decides ADR 0037 and is separately authorized.

APG64 Markdown architecture peer review

APG64 delivers and independently reviews exact APG63. It reproduces seven material architecture defects, applies one coherent forward correction, preserves exact corrected-state evidence, and accepts ADR 0037 with amendment after fresh 36/36 replay and source/target review. The accepted architecture remains authoring-eligible-with-narrowing, but no Markdown skill is authored or integrated. Development remains 28/28/28 with 14 stable / 14 provisional; CSS stays absent; public, active, and target state is unchanged; and no successor is authorized.

APG65 Markdown language-profile candidate authoring

APG65 authors one fresh branch-only markdown-language-profile candidate from the accepted ADR 0037 architecture: leaf, candidate specification, and navigation-only scenario-coverage record, with ADR 0038 Proposed. The candidate is branch-only, Proposed, unintegrated, and pending separately authorized APG66 validation; it contains no numeric whole-file band and creates no current integration owner. Integrated development remains 28/28/28 with 14 stable / 14 provisional; CSS stays absent; public, active, and target state is unchanged; and no successor is authorized.

APG66 Markdown language-profile validation and integration

APG66 delivers exact APG65, independently replays the accepted 34-scenario oracle, freezes one coherent correction before fresh non-author review, and accepts ADR 0038 with amendment. The profile is retained provisionally with every current owner: development is 29/29/29, 14 stable / 15 provisional, with 27 general routes, one ChatGPT-local route, and 28 checked edges. ADR 0037 and APG65 authorship are preserved. Public and active corrected v0.4.0 and the read-only targets remain unchanged; no readiness, publication, deployment, APG67, or successor begins.

APG67 JavaScript language-profile architecture

APG67 proposes the JavaScript language-profile architecture and candidate-independent lean validation contract under ADR 0039 (Proposed): the exact ECMAScript 2026 annual source is the stable reference, JavaScript, TypeScript, and Node.js remain separate owners, structural policy is qualitative with no numeric whole-file bands, and a frozen forty-row register binds future authoring and review. No JavaScript skill exists; development remains 29/29/29 with 14 stable / 15 provisional and 27/1/28 routes; Markdown remains retained provisional; corrected public and active v0.4.0 and the read-only targets are unchanged. APG68 is recommended but not begun.

APG68 JavaScript architecture peer review

APG68 normally delivers and preserves exact APG67, independently reconstructs and replays the JavaScript architecture, and applies one coherent correction. ADR 0039 is Rejected after fresh review finds material defects in the sole corrected state; eligibility is not-applicable-rejected. No JavaScript skill is authored or integrated; development remains 29/29/29, 14/15, and 27/1/28. APG69 is not recommended or begun.

APG69 JavaScript core architecture reset

APG69 exercises new human authority to propose ADR 0040 (Proposed): a narrower ECMAScript-core owner with question-specific typed authorities, a four-layer semantic/structural/policy/effective decision model with an ordered effective route union, and independent semantic, structural, composition, and process registers. ADR 0039 remains Rejected and unchanged; no JavaScript skill is authored or integrated; development remains 29/29/29, 14/15, and 27/1/28. Separately authorized APG70 terminally decides ADR 0040.

APG70 JavaScript core layered-architecture peer review

APG70 normally delivers exact APG69, independently reproduces the initial material defects, applies one coherent correction, and preserves its actual patch. Fresh non-author review then finds new material source-identity, signal-predicate, context-route, policy-typing, checker, and owner-binding defects. ADR 0040 is Rejected; eligibility is not-applicable-rejected; no current JavaScript architecture input, skill, or integration owner exists. Development remains 29/29/29, 14/15, and 27/1/28. APG71 is not recommended or begun.

APG71 TypeScript architecture and compiler-generation boundary

APG71, under separate new human authority for TypeScript architecture only, proposes ADR 0041 (Proposed): a narrow TypeScript static-semantics owner under project-selected exact compiler authority, a first-class TypeScript 6/7 compiler-generation boundary, closed source-kind and embedded-host boundaries, structural disposition D, and eligibility authoring-eligible-with-narrowing with no candidate-authoring authority. ADR 0039 and ADR 0040 remain Rejected; no TypeScript skill or integration owner exists; development remains 29/29/29, 14/15, and 27/1/28. Separately authorized APG72 terminally decides ADR 0041.

APG72 TypeScript architecture peer review

APG72 preserves exact APG71, independently rebuilds the oracle, freezes the complete initial set, and applies one coherent correction. Two fresh non-author lanes find new material owner/route, role-state, source-kind, and evidence-state defects. ADR 0041 is Rejected; eligibility is not-applicable-rejected; no current TypeScript architecture input or skill exists. Development remains 29/29/29, 14/15, and 27/1/28. APG73 is not recommended or begun.

APG73 language-profile production recovery charter

APG73 records new human product authority and accepts ADR 0042. The controlling production recovery charter and iterative hardening contract replace automatic rejection after one correction with up to three separately evidenced hardening rounds by default. Critical and High defects still block integration; explicit human acceptance is required for Medium or Low debt and for terminal rejection or removal. TypeScript is essential, CSS and JavaScript are desirable, JSX is deferred, TypeScript 7 is the intended primary generation, and temporary TypeScript 6 is permitted only for an exact required role. No profile is authored or integrated. Development stays 29/29/29, 14/15, and 27/1/28; Markdown, rejected ADRs, public/active v0.4.0, and targets remain unchanged. APG74 is recommended but not begun.

APG74 TypeScript language-profile candidate

APG74 authors the first ADR 0042 product-recovery candidate: the TypeScript language-profile leaf, its candidate specification and scenario coverage, and the TypeScript 7 intended-state fixture, proposing ADR 0043. The exact stable compiler is freshly selected (typescript@7.0.2); the TypeScript 6 compatibility disposition is not-required with a recorded refresh condition; and the freshly pinned targets show the live theme checking under typescript@5.9.3 — migration baseline, not the destination. The candidate is authored-proposed-unintegrated: the APG74 branch carries a transitional 30/29/29 shape while integrated development remains exact APG73 at 29/29/29, 14/15, and 27/1/28. Markdown, rejected ADRs, public/active v0.4.0, and targets remain unchanged. APG75 owns iterative hardening and the terminal ADR 0043 decision; it is recommended but not begun.

APG75 TypeScript iterative hardening and provisional integration

APG75 normally delivers exact APG74, independently rebuilds its 24 semantic and 14 fixture vectors, and uses three separately preserved hardening rounds. The final exact-TypeScript-7 suite passes 78 focused tests and fresh review finds zero Critical, High, Medium, or Low defects. ADR 0043 is Accepted with amendment and typescript-language-profile is retained provisionally. Development is 30/30/30, 14/16, and 28/1/29. A disposable rollback restores 29/29/29, 14/15, and 27/1/28 while preserving history, decisions, Markdown, and corrected public and active v0.4.0. APG75 performs no target mutation, readiness, publication, deployment, stable-maturity, APG76, or successor work.

APG75A TypeScript scope and lifecycle closure

APG75A closes five post-integration defects without rewriting APG74 or APG75. The reusable profile is compiler-generation-neutral and consumes exact project-selected evidence; Theme Forge Terminal Nova retains its separate TypeScript 7 product destination. Response has exactly four values and remains disjoint from Selection. Coverage, fixture, manifest, and parser lifecycle now agree on provisional integration, and standard runners fail early unless exact TypeScript 7.0.2 is explicitly bound. ADR 0043 remains Accepted with amendment; development remains 30/30/30, 14/16, and 28/1/29; public/active corrected v0.4.0 and targets are unchanged. APG76 is recommended but not begun.

APG76 CSS language-profile candidate recovery

APG76 proposes ADR 0044 on the Claude authoring branch: one reusable CSS semantics profile, one navigation-only twenty-four-scenario coverage record, and one fourteen-case target-first fixture. CSS remains unintegrated, with no catalog row, projection, maturity row, route, project selection, release owner, or maintained test owner. TypeScript and Markdown remain retained provisional and JavaScript remains absent. Published and active corrected v0.4.0 and both read-only targets are unchanged and unexecuted. APG77 Codex CSS iterative hardening is recommended but not begun.

APG77 CSS iterative hardening repair checkpoint

APG77 preserves three immutable correction rounds and the APG76 candidate. Fresh terminal review finds two High retained-evidence defects, so ADR 0044 remains Proposed and CSS remains unintegrated. Development stays 30/30/30, 14/16, and 28/1/29; public and active corrected v0.4.0 and both read-only targets are unchanged. Further repair requires a human continuation decision.

APG77A CSS evidence-retention repair checkpoint

APG77A preserves the explicitly authorized target-identity and independent-vector correction. Fresh review finds three High defects, including a wrong SVG authority, false-pass H2 qualification, and copied target bytes. ADR 0044 stays Proposed and CSS remains unintegrated at 30/30/30, 14/16, and 28/1/29. Main, corrected public/active v0.4.0, and both targets are unchanged. APG78 is not recommended; a new human decision is required.

APG77B CSS traceability and clean-room repair checkpoint

APG77B preserves the explicitly authorized H3-H5 correction. Fresh four-lane review finds four High and two Medium defects in qualification, generated-report and escaped-copy boundaries, immutable diff binding, and artifact proportionality. No debt is accepted. ADR 0044 stays Proposed and CSS remains unintegrated at 30/30/30, 14/16, and 28/1/29. Main, corrected public/active v0.4.0, and both targets are unchanged. APG78 is not recommended; a new human decision is required.

APG77C CSS evidence-proportionality repair checkpoint

APG77C preserves the human-authorized P1-P6 correction, makes exact private deletion patches a narrow historical exception, and replaces exhaustive current prose provenance with compact consequence-bearing evidence. Fresh review finds six Medium and one Low unaccepted qualification defects. ADR 0044 stays Proposed and CSS remains unintegrated at 30/30/30, 14/16, and 28/1/29. Main, corrected public/active v0.4.0, provisional TypeScript and Markdown, and both targets are unchanged. APG78 is not recommended; a new human decision is required.

APG77D CSS known debt and provisional integration

APG77D records explicit human acceptance of exactly four Medium and one Low qualification limitations, accepts no Critical/High or semantic/source/target/ runtime/release/rollback debt, and closes proportionality, live lifecycle, release, historical-exclusion, and disposable-rollback gates. ADR 0044 is Accepted with amendment and CSS is provisionally integrated with known debt at 31/31/31, 14/17, and 29/1/30. TypeScript and Markdown remain provisional; public/active corrected v0.4.0 and both unexecuted targets remain unchanged. APG78 is recommended but not begun.

APG79 JavaScript repair checkpoint

APG79 preserves three correction rounds and independently validates the APG78 candidate, fixture, exact engine, sources, rights, and target bindings. Terminal review finds zero Critical, two High, two Medium, and zero Low material defects, with no accepted JavaScript debt. ADR 0045 remains Proposed and JavaScript is repair-required-after-round-3 without integration. Development remains exact APG77D at 31/31/31, 14/17, and 29/1/30; CSS known debt, provisional CSS, TypeScript and Markdown, corrected public/active v0.4.0, and targets remain unchanged. APG80 is not recommended; a human continuation decision is required.

APG79A JavaScript terminal-proof repair checkpoint

APG79A preserves one separately authorized immutable correction for APG79's four terminal findings. Four fresh non-author lanes find five Medium material defects and no Critical, High, or Low finding; no JavaScript debt is accepted. ADR 0045 remains Proposed and JavaScript is repair-required-after-apg79a without integration. Development remains exact APG77D at 31/31/31, 14/17, and 29/1/30; CSS known debt, provisional CSS/TypeScript/Markdown, corrected public/active v0.4.0, and both targets remain unchanged. APG80 is not recommended and a new human decision is required.

APG79B JavaScript contract and harness repair checkpoint

APG79B preserves one separately authorized immutable correction for APG79A's five Medium findings. Five fresh non-author lanes find four unique Medium material defects and no Critical, High, or Low finding; no JavaScript debt is accepted. ADR 0045 remains Proposed and JavaScript is repair-required-after-apg79b without integration. Development remains exact APG77D at 31/31/31, 14/17, and 29/1/30; CSS debt, public/active corrected v0.4.0, and both targets remain unchanged. APG80 is not recommended and a new human decision is required.

APG79C JavaScript human-debt decision and integration checkpoint

APG79C accepts exactly four Medium JavaScript qualification limitations for provisional use, with zero Critical or High and no semantic, source, target, owner, release, or rollback debt accepted. A separate unaccepted Medium Test262 identity discrepancy blocks integration. ADR 0045 remains Proposed and JavaScript remains repair-required-after-apg79b and unintegrated. Development stays exact APG77D at 31/31/31, 14/17, and 29/1/30; APG80 is not recommended.

APG79E JavaScript report-binding debt and provisional integration

APG79E accepts JS-QD-005 as one additional Medium supporting qualification limitation and directly verifies the current APG79B report rather than claiming the maintained proxy is repaired. With zero Critical, High, or unaccepted Medium/Low findings and every product gate green, ADR 0045 is Accepted with amendment and JavaScript is provisionally integrated under exactly JS-QD-001 through JS-QD-005. Development is 32/32/32, 14/18, and 30/1/31. Public and active corrected v0.4.0 and both read-only targets remain unchanged; APG80 is recommended but not begun. APG80 authors a narrow reusable nodejs-runtime-profile candidate with a twenty-four-scenario navigation record, a fourteen-case APG-owned target-first fixture, and Proposed ADR 0046. The candidate consumes runtime roles rather than selecting them, because fresh target evidence shows the exact Node version is never resolvable from project evidence alone in either read-only target. Nothing is integrated: mainline development remains 32/32/32, 14/18, and 30/1/31, known debt remains exactly CSS-QD-001 through CSS-QD-005 and JS-QD-001 through JS-QD-005, and public and active corrected v0.4.0 and both targets are unchanged. APG81 is recommended but not begun.

APG81 Node.js hardening repair checkpoint

APG81 preserves three immutable correction rounds and completes the required terminal review. The final ledger is zero Critical, three High, two Medium, and zero Low; no Node debt is accepted. ADR 0046 remains Proposed and nodejs-runtime-profile remains branch-only, repair-required-after-round-3, and unintegrated. Main remains exact APG79E at 32/32/32, 14/18, and 30/1/31; the candidate branch remains 33/32/32. Existing CSS/JavaScript debt, corrected public/active v0.4.0, and both unexecuted targets are unchanged. A new human continuation decision is required.

APG81A Node.js threat-model correction checkpoint

APG81A preserves a zero-finding correction that retires immutable-flag and copied- runtime sealing claims in favor of a controlled local-or-CI qualification contract with direct runtime pre/post observation. Terminal integration review finds one High rollback defect and one Medium contradictory ADR-index lifecycle defect in the proposed integration state, so those bytes are not retained. ADR 0046 remains Proposed, Node remains corrected, repair-required, and unintegrated with zero Node debt, and main remains 32/32/32, 14/18, and 30/1/31. A new human decision is required; no successor is authorized.

APG81B Node.js integration-contract clarification checkpoint

APG81B clarifies that candidate-preserving rollback targets 33/32/32. Review of the complete reconstructed integration candidate finds three Medium defects and no Critical, High, or Low finding, with zero Node debt. The candidate's rollback lifecycle was contradictory, its pending review records contradicted its proposed message, and one test-only xdist repair exceeded integration authority. The attempted integration is discarded. ADR 0046 remains Proposed and Node remains corrected, repair-required, and unintegrated at branch shape 33/32/32; main remains 32/32/32, 14/18, and 30/1/31. A new human decision is required.

APG81C Node.js lifecycle, test, and scratch repair checkpoint

APG81C retains the integration-support correction for rollback lifecycle, external review sequencing, per-invocation xdist cleanup, and ignored machine-local scratch. Fresh precommit review passes with zero findings and zero Node debt, but immutable review finds one Medium contradiction in retained pending-review prose. Integration stops. ADR 0046 remains Proposed, Node remains corrected, repair-required, and unintegrated at 33/32/32, and exact APG79E main remains 32/32/32, 14/18, and 30/1/31. No successor is authorized.

APG81D Node.js repo-local scratch integration checkpoint

APG81D reconstructs the complete Node provisional-integration candidate and temporary ignored repo-local scratch contract. Final staged review finds one High rollback test- contract defect and three Medium specification, fail-closed query, and cleanup- evidence defects. The attempted integration is discarded without product repair. ADR 0046 remains Proposed; Node remains repair-required and unintegrated at branch shape 33/32/32; exact APG79E main remains 32/32/32, 14/18, and 30/1/31. No Node or scratch debt is accepted and no successor is authorized.

APG81E Node.js final-integration review checkpoint

APG81E reconstructs the complete Node integration candidate and closes the APG81D scratch-cleanup proof before review. Final staged review nevertheless finds one High incomplete rollback test-owner projection and two Medium chronology and fresh release-schedule evidence defects. The attempted integration is discarded. ADR 0046 remains Proposed; Node remains repair-required and unintegrated at branch shape 33/32/32; exact APG79E main remains 32/32/32, 14/18, and 30/1/31. No Node or scratch debt is accepted and no successor is authorized.

APG81F Node.js actual-test-projection repair checkpoint

APG81F attempts the one authorized actual-test-projection, chronology, and source- evidence correction. Fresh review finds three High and two Medium defects, so all attempted runner, test, and correction bytes are discarded. ADR 0046 remains Proposed; Node remains retained, repair-required, and unintegrated at 33/32/32. Exact APG79E main and all public, active, target, and debt state remain unchanged. No successor is authorized.

APG81G Node.js selector, release, and integration closure

APG81G attempted the one authorized correction of APG81F's five findings. Fresh review found two High and four Medium defects, so all attempted implementation and test bytes are discarded. ADR 0046 remains Proposed and Node remains retained, repair-required, and unintegrated at 33/32/32. No Node or scratch debt is accepted, and no successor is authorized.

APG81H Node.js reviewable qualification and integration closure

APG81H is the explicit human continuation after APG81G. Fresh review of its one correction found that the frozen topology evidence did not bind the complete actual language-profile lifecycle map. The attempted implementation and test bytes were discarded before commit. ADR 0046 remains Proposed and Node remains retained, repair-required, and unintegrated at 33/32/32. No integration, rejection, removal, debt acceptance, or successor work occurred.

The resumed APG81H recovery isolates the generic change-size negative in a real bare-repository fixture and provisionally integrates the retained Node.js profile. Development is 33/33/33, 14/19, and 31/1/32; Node lifecycle is provisionally-integrated and ADR 0046 is Accepted with amendment. Node and scratch debt are zero. Public and active corrected v0.4.0 remain unchanged, and no stable maturity, readiness, publication, deployment, APG81I, APG82, or successor work is implied.

APG82 APGR CLI and distribution foundation

APG82 establishes the publishable agentic-praxis-grimoire Python distribution, the agentic_praxis_grimoire import package, and the canonical apgr executable. Installed consumer commands use a packaged, source-digest-bound skill-metadata manifest without an APG source checkout. Full skill content and projection remain repository-maintenance operations, which fail clearly when repository authority is absent. Maintained historical executable names remain thin compatibility shims over the same owners.

APGR resolves --apgr-home over APGR_HOME over ~/.apgr. Declarative global and project configuration use config.toml beneath the resolved APGR home and <project-root>/.apgr/, with scalar precedence of explicit CLI, project, global, then built-in default. The reserved future adapter checkout is ~/.apgr/agentic-praxis-grimoire-nd/; APG82 creates no adapter or Nix deployment.

New terminal artifacts use ~/Documents/agent/outbox/<project>/<phase>/: exactly one current Git-show, Git-diff, or ops-only primary plus immutable numbered final responses. The apgr skills context-report measurement reports exact discoverable-description bytes and characters without imposing a budget threshold. APG82 publishes nothing and starts no successor. APG83 bounded dogfood and release readiness is recommended next under separate authority.

APG83 v0.5 release readiness

APG83 dogfoods the retained v0.5 product against exact revisions of the two approved Knowledge Forge repositories and qualifies APGR from fresh outside-checkout installs. The compact matrix produces eleven passes, one truthful deferred-v0.6 boundary, and no material APG blocker. One deterministic sdist archive-metadata defect is corrected and covered before the final release rebuild.

APG83 proves the v0.5 distribution contract: the public Git release carries the complete skill corpus, projections, and maintenance owners, while the PyPI distribution carries the checkout-independent apgr runtime and exact skill metadata. The pair is version-bound and reproducibly consumable without mutable development-checkout state. APG84 publication is recommended but separately authorized and is not begun here.