@sabaiway/agent-workflow-memory 2.0.0 → 2.2.0

package/CHANGELOG.md

@@ -4,6 +4,30 @@ All notable changes to the memory substrate. Versions are this **package's** npm
 they are distinct from the **deployment-lineage** stamp written into a project's
 `docs/ai/.memory-version` (which tracks the shared `agent-workflow` lineage, head `2.0.0`).
 
+## 2.2.0 — agent_rules template re-render: the lens gains the prompt-economy clause (REC-UX-REWORK D7, AD-053)
+
+A **feature** release (ships with kit 1.46.0 / engine 1.16.0; the deployment lineage head stays
+`2.0.0` — no deployed-`docs/ai` structure change). The bundled
+`references/templates/agent_rules.md` §2.6 lens block is re-rendered to the current engine canon:
+its cost-lanes line now carries the **prompt-economy clause** (read-only fan-out on restricted-tool
+vehicles only; one plain pipeline per call; capability-gated launcher guidance; judgment, code and
+synthesis stay at the frontier lane; the stated honest limit). Template-only — seeding, upgrade and
+every other substrate behavior are unchanged; the render is byte-parity-pinned against the engine
+fragment by the kit's lens-mirror guard.
+
+## 2.1.0 — Autonomy template seed (AD-044 Plan 4)
+
+A **feature** release (ships with kit 1.45.0 / engine 1.15.0; the deployment lineage head stays
+`2.0.0` — the seed is ensure-if-missing, stamp-independent). `references/templates/autonomy.json`
+— the sparse, defaults-equivalent autonomy-policy seed (`_README` only): resolving it equals
+resolving no file at all, TEST-PINNED, so the seed can never drift from the computed defaults
+(commit/push/publish `ask`; network/credentials/fs-outside-repo `deny`; absent activities floor at
+`prompt`); it is STRUCTURALLY a seed (meta keys only) — an explicit policy declaring the same
+values reads as a real declaration on every kit surface. Lifecycle follows gates.json: bootstrap
+seeds it, upgrade ensures-if-missing (byte-preserving), hand-editable; the kit mirrors it via the
+template manifest. Tarball sentinel 43→44. Deliberately NOT a delegated-bootstrap required asset
+(the AD-044 Plan-3 no-gate decision holds).
+
 ## 2.0.0 — One-file-per-ADR store: the 3-tier decisions cascade retired (AD-051)
 
 A **MAJOR** release (BREAKING; co-released with the workflow kit 1.42.0). The deployment-lineage

package/README.md

@@ -33,11 +33,12 @@ Open your agent inside a project and run the skill:
 - **`/agent-workflow-memory`** — bootstrap a new or empty project. Asks the three setup
   questions (visibility, conversational language, agent attribution), then writes `docs/ai/`,
   `AGENTS.md`, the enforcement scripts + pre-commit hook, and stamps
-  `docs/ai/.memory-version`. The two pointer slots in `AGENTS.md` (the **methodology** and
-  **orchestration** pointers, filled later by the family composition root) are left **empty**.
+  `docs/ai/.memory-version`. The three pointer slots in `AGENTS.md` (the **methodology**,
+  **orchestration** and **autonomy** pointers, filled later by the family composition root) are
+  left **empty**.
 - **`/agent-workflow-memory upgrade`** — migrate an existing deployment to the skill's current
   version. Reads `docs/ai/.memory-version`, runs only newer migrations in semver order, and
-  **preserves** anything already in either pointer slot (extract-and-reinsert — it never
+  **preserves** anything already in every pointer slot (extract-and-reinsert — it never
   regenerates `AGENTS.md` wholesale).
 
 ## Stamps & lineage

package/SKILL.md

@@ -3,7 +3,7 @@ name: agent-workflow-memory
 description: Deploy or upgrade a portable AI-agent memory substrate in any project — an entry-point `AGENTS.md` (+ `CLAUDE.md` alias) and a structured `docs/ai/` context store with cap/archive/index enforcement. Use when the user wants to bootstrap `docs/ai/`, set up the Memory Map and session protocols, install the docs-rotation pre-commit hook, or run `/agent-workflow-memory` / `/agent-workflow-memory upgrade`. Triggers on "set up the memory system", "deploy the AI memory here", "bootstrap docs/ai", "upgrade the memory substrate". This is the substrate only — the workflow methodology (plan→execute→review, queue, Cleanup) is owned elsewhere and injected into AGENTS.md by the family composition root.
 disable-model-invocation: true
 metadata:
-  version: '2.0.0'
+  version: '2.2.0'
 ---
 
 # agent-workflow-memory
@@ -20,12 +20,13 @@ manually).
 
 This skill is the **memory layer** of the `agent-workflow` family. It **knows nobody else** in
 the family. In particular it does **not** own the **workflow methodology** (plan → execute →
-review vocabulary, lifecycle, `docs/plans/queue.md`, mandatory Cleanup, plan-then-execute) **or the
-orchestration recipes** — those are injected into **two delimited pointer slots** in `AGENTS.md`
-(`workflow:methodology` + `workflow:orchestration`) by the family **composition root**, never by this
+review vocabulary, lifecycle, `docs/plans/queue.md`, mandatory Cleanup, plan-then-execute), **the
+orchestration recipes, or the autonomy-policy contract** — those are injected into **three delimited
+pointer slots** in `AGENTS.md` (`workflow:methodology` + `workflow:orchestration` +
+`workflow:autonomy`) by the family **composition root**, never by this
 skill. (This substrate does not name or depend on any specific sibling — it only honours the slot
-contract.) This skill only ever ships **both** slots **empty** and **preserves** whatever is already
-in them on upgrade.
+contract.) This skill only ever ships **all three** slots **empty** and **preserves** whatever is
+already in them on upgrade.
 
 The substrate **artifacts** (this skill, the templates, the deployed `docs/ai/` files) stay in
 their **source language** — for cross-agent and cross-team portability. That is separate from the
@@ -36,8 +37,8 @@ contents).
 
 ## Ownership table
 
-What this substrate owns vs what it only points at. The methodology + orchestration pointers are the
-**two empty slots** the composition root fills — never author that text here.
+What this substrate owns vs what it only points at. The methodology + orchestration + autonomy
+pointers are the **three empty slots** the composition root fills — never author that text here.
 
 | Concern | Owner | In the deployed `AGENTS.md` |
 |---|---|---|
@@ -47,9 +48,11 @@ What this substrate owns vs what it only points at. The methodology + orchestrat
 | Deployment-lineage stamp | **memory** | `docs/ai/.memory-version` |
 | Plan→Phase→Step vocabulary, lifecycle, `queue.md`, mandatory Cleanup | **methodology** (not this skill) | the empty `workflow:methodology` slot — filled by the composition root |
 | Orchestration recipes (Solo / Reviewed / Council / Delegated) | **methodology engine** (not this skill) | the empty `workflow:orchestration` slot — filled by the composition root |
+| Autonomy-policy read contract (`docs/ai/autonomy.json`) | **methodology engine** (not this skill) | the empty `workflow:autonomy` slot — filled by the composition root |
 | Per-project recipe **CONFIG** (which recipe each activity/slot uses) | **memory** seeds an *editable default* | `docs/ai/orchestration.json` (agent-writable via the composition root's `set-recipe` writer, or hand-edited; the recipe **canon** + the slot **vocabulary** live in the engine / composition root, never here) |
 | Per-project **gate declaration** (which verification commands must be green) | **memory** seeds an *editable default* | `docs/ai/gates.json` (hand-editable; an empty list as shipped — the project declares its own commands; the **runner** lives in the composition root, never here) |
 | Per-project **verification profile** (optional; the fold-completeness language-independence contract) | **memory** seeds an *editable default* | `docs/ai/verification-profile.json` (hand-editable; the seeded default reproduces the composition root's default V8 + node:test behaviour — delete it for exactly that; declares coverage source / single-test format / optional SARIF path; the **runner** lives in the composition root, never here) |
+| Per-project **autonomy policy** (red-lines + per-activity autonomy level) | **memory** seeds an *editable default* | `docs/ai/autonomy.json` (hand-editable, or agent-writable via the composition root's `set-autonomy` writer; the seed is SPARSE — the onboarding note only, defaults-equivalent, so deploying it never changes behavior; the policy **canon** + the render live in the engine / composition root, never here) |
 
 ---
 
@@ -97,19 +100,22 @@ bootstrapping over a live system, but the user makes the final call.
 5. **Entry-point doc.** If `AGENTS.md` / `CLAUDE.md` already exist (recon), do **not**
    overwrite — show the user and ask whether to merge or replace. Otherwise create `AGENTS.md`
    from `${CLAUDE_SKILL_DIR}/references/templates/AGENTS.md` and symlink `CLAUDE.md -> AGENTS.md`
-   (`ln -s AGENTS.md CLAUDE.md`). **Leave BOTH pointer slots (`workflow:methodology` +
-   `workflow:orchestration`) exactly as shipped — empty.** Filling them is the composition root's job.
+   (`ln -s AGENTS.md CLAUDE.md`). **Leave ALL THREE pointer slots (`workflow:methodology` +
+   `workflow:orchestration` + `workflow:autonomy`) exactly as shipped — empty.** Filling them is the
+   composition root's job.
 6. **Deploy `docs/ai/`.** Create the files + `pages/` + the seed `adr/` store from
    `${CLAUDE_SKILL_DIR}/references/templates/` (every non-`AGENTS.md`, non-`adr-record.md` template —
    the latter is a skill-home authoring reference, never deployed). This deploys the HOT ADR window
    `docs/ai/decisions.md` **and** the seed navigator `docs/ai/adr/log.md` (its `adr/` subdir copies
-   in wholesale), plus the three seeded, **user-editable** strict-JSON configs:
+   in wholesale), plus the four seeded, **user-editable** strict-JSON configs:
    `docs/ai/orchestration.json` — the per-project recipe defaults the composition root's `procedures`
    advisor reads — `docs/ai/gates.json` — the project's gate declaration, an empty list to fill with
-   its own verification commands, consumed by the composition root's gate runner — and
+   its own verification commands, consumed by the composition root's gate runner —
    `docs/ai/verification-profile.json` — the OPTIONAL fold-completeness verification profile (the
    language-independence contract; the seeded default reproduces the composition root's default V8 +
-   node:test behaviour, delete it for exactly that). Keep each `.md` file's frontmatter.
+   node:test behaviour, delete it for exactly that) — and `docs/ai/autonomy.json` — the per-project
+   autonomy policy, seeded SPARSE (the onboarding note only, defaults-equivalent — behavior changes
+   only when the user declares levels). Keep each `.md` file's frontmatter.
 7. **Fill templates** per the table below.
 8. **Install enforcement (Node projects).** Copy `${CLAUDE_SKILL_DIR}/references/scripts/*.mjs`
    (+ `*.test.mjs`) into the project's `scripts/`. **No Node runtime** → skip this + the hook;
@@ -136,7 +142,7 @@ bootstrapping over a live system, but the user makes the final call.
 > upgrade**, do the write steps (bootstrap 1–10 / upgrade 1–7: write `docs/ai/` + `AGENTS.md` +
 > `.memory-version`, **including seeding / stamp-independently ensuring `docs/ai/orchestration.json`**)
 > but **do NOT** run **any** commit gate and **do NOT** ask to commit — the
-> composition root owns the **single** commit gate, raised after it injects the two pointer slots.
+> composition root owns the **single** commit gate, raised after it injects the three pointer slots.
 > The three setup answers and the target dir are passed in by the composition root; you perform no
 > commit and no slot injection. **Standalone** invocation keeps its own commit gate (bootstrap step
 > 11 / upgrade step 7).
@@ -178,8 +184,9 @@ Fill strategy:
    **untracked AND not ignored** → **AMBIGUOUS** → **ASK** the user before writing. This visibility check
    runs on **every** in-range upgrade, even at head — it is not gated by the stamp delta, but it is gated
    **behind** the never-downgrade STOP above. **Also stamp-independent (same gate, before the equal-head
-   short-circuit): ensure the THREE seeded `.json` configs** — for `docs/ai/orchestration.json`,
-   `docs/ai/gates.json`, AND `docs/ai/verification-profile.json`, **create the file from its
+   short-circuit): ensure the FOUR seeded `.json` configs** — for `docs/ai/orchestration.json`,
+   `docs/ai/gates.json`, `docs/ai/verification-profile.json`, AND `docs/ai/autonomy.json`,
+   **create the file from its
    `${CLAUDE_SKILL_DIR}/references/templates/` template if missing**, **preserve it byte-for-byte
    if it already exists** (a user may have edited it; never clobber it). The shipped
    `orchestration.json` template's `_README` already frames that config as agent-writable (set it
@@ -187,7 +194,9 @@ Fill strategy:
    declaration (what to verify — consumed by the composition root's gate runner);
    `verification-profile.json` is the OPTIONAL fold-completeness verification profile — the seeded
    default reproduces the composition root's default V8 + node:test behaviour, so a project deletes
-   it to fall back to exactly that, or edits it for another language/runner. (Refreshing the
+   it to fall back to exactly that, or edits it for another language/runner; `autonomy.json` is the
+   per-project autonomy policy, seeded SPARSE (the onboarding note only, defaults-equivalent —
+   agent-writable via the composition root's `set-autonomy` writer, hand-editing supported). (Refreshing the
    orchestration `_README` note in place on an existing file is the **composition root's** job on
    its own reconcile — this substrate is standalone and only seeds-or-preserves; it owns no
    cross-package refresh helper. The gates declaration gets no note-refresh at all — authored
@@ -232,12 +241,13 @@ Fill strategy:
    (`migrations/1.1.0-communication-language.md`); pre-1.2.0 with no *Attribution* block, ask +
    insert defaulting to `off` (`migrations/1.2.0-agent-attribution.md`). (An answer already
    collected by the step-4 batched prompt is carried in — never re-asked here.)
-6. **Preserve BOTH pointer slots.** If `AGENTS.md` has the `workflow:methodology` and/or
-   `workflow:orchestration` markers, **never regenerate the file wholesale** — extract any bytes
-   between each pair and reinsert them unchanged. If a pair is absent (a legacy `AGENTS.md`),
-   gracefully **no-op** on that slot (adding a slot to already-deployed files is the composition
-   root's reconcile, not this substrate's job). On any malformed marker state (single, reversed,
-   nested, or duplicate pair), **no-op with an error** — never edit.
+6. **Preserve ALL pointer slots.** If `AGENTS.md` has the `workflow:methodology`,
+   `workflow:orchestration` and/or `workflow:autonomy` markers, **never regenerate the file
+   wholesale** — extract any bytes between each pair and reinsert them unchanged. If a pair is
+   absent (a legacy `AGENTS.md`), gracefully **no-op** on that slot (adding a slot to
+   already-deployed files is the composition root's reconcile, not this substrate's job). On any
+   malformed marker state (single, reversed, nested, or duplicate pair), **no-op with an error** —
+   never edit.
 7. **Re-stamp** `docs/ai/.memory-version` to the lineage head (atomic write — mechanics unchanged).
    Report changes **in plain human terms** (which parts of the deployment are now different);
    **omit the raw structure number**, and do not recite the two-axes note here (it belongs to
@@ -279,9 +289,9 @@ memory substrate has none, so it relies on the STOP + the explicit ask — never
 - **Stamp = lineage head, not package version.** `.memory-version` carries the **deployment-lineage
   head** (`2.0.0`, the shared `agent-workflow` lineage) — a **separate axis** from this package's npm
   version (the two may even coincide by accident). They move independently.
-- **Both pointer slots ship empty and stay the user's.** Never author methodology or orchestration
-  text into them; on upgrade, preserve their content byte-for-byte. The composition root is their only
-  writer.
+- **All three pointer slots ship empty and stay the user's.** Never author methodology,
+  orchestration or autonomy text into them; on upgrade, preserve their content byte-for-byte. The
+  composition root is their only writer.
 - **The `Co-Authored-By` trailer is added by the harness, not by prose.** When attribution is
   `off` + Claude Code, also set `"includeCoAuthoredBy": false` in `.claude/settings.json`.
 - **Hidden mode is project-local, never machine-global.** Memory's hide writes to the repo's
@@ -307,10 +317,10 @@ The three setup choices each have a full contract in
 ## References
 
 - [`references/contracts.md`](references/contracts.md) — the three setup contracts in full.
-- [`references/templates/`](references/templates/) — stack-agnostic `AGENTS.md` (with the two empty
-  pointer slots — methodology + orchestration), `agent_rules.md`, the seeded user-editable
-  `orchestration.json` config, the `adr-record.md` ADR authoring reference + the seed `adr/log.md`
-  navigator, and all `docs/ai/` files to deploy.
+- [`references/templates/`](references/templates/) — stack-agnostic `AGENTS.md` (with the three empty
+  pointer slots — methodology + orchestration + autonomy), `agent_rules.md`, the seeded user-editable
+  `orchestration.json` + `autonomy.json` configs, the `adr-record.md` ADR authoring reference + the
+  seed `adr/log.md` navigator, and all `docs/ai/` files to deploy.
 - [`references/scripts/`](references/scripts/) — the Node enforcement scripts (caps + staleness +
   index-freshness gate, one-file-per-ADR archive, hook installer) and their unit tests.
 - [`scripts/stamp-takeover.mjs`](scripts/stamp-takeover.mjs) — the upgrade-time lineage state

package/capability.json

@@ -3,7 +3,7 @@
   "schema": 1,
   "name": "agent-workflow-memory",
   "kind": "memory-substrate",
-  "version": "2.0.0",
+  "version": "2.2.0",
   "provides": ["context"],
   "roles": {},
   "detect": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sabaiway/agent-workflow-memory",
-  "version": "2.0.0",
+  "version": "2.2.0",
   "description": "Portable, cross-agent memory substrate for AI coding agents — an AGENTS.md entry point + docs/ai context with cap/archive/index enforcement, deployable standalone or as part of the agent-workflow family. The memory layer of the agent-workflow family.",
   "keywords": [
     "ai-agents",

package/references/templates/AGENTS.md

@@ -57,6 +57,8 @@ Start-of-session, during-work, and task-completion procedures live in [`docs/ai/
 <!-- workflow:methodology:end -->
 <!-- workflow:orchestration:start -->
 <!-- workflow:orchestration:end -->
+<!-- workflow:autonomy:start -->
+<!-- workflow:autonomy:end -->
 
 ---
 
@@ -68,7 +70,6 @@ Start-of-session, during-work, and task-completion procedures live in [`docs/ai/
 | No `any` / no unsafe casts | Linter / type-checker |
 | Functional style (no mutation in app code) | Linter |
 | Interactive elements semantic (button/link, not div+onClick) | Linter / a11y |
-| No hardcoded colors — design tokens only | Code review |
 | No business logic in components → hooks/services | Architecture review |
 | No changes without tests (TDD) | Required |
 | Check page docs before changes; update them after | Process |

package/references/templates/agent_rules.md

@@ -85,7 +85,7 @@ Apply these when authoring a plan, reviewing, folding a finding, or editing code
 - **Per-round emission.** Every review round emits **{round N · finding-origin tally · per-backend verdict}** so the crossover is a computed, visible signal, not a remembered rule.
 - **Recipe fidelity.** Council runs every backend the recipe names, **every round**; silently dropping a ready backend for quota/convenience is a forbidden downgrade — an unavailable backend is a LOUD, stated degrade, never a quiet drop.
 - **ExitPlanMode ≠ execute.** A harness "approved — start coding" prompt authorizes the PLAN only; this methodology overrides it. Continue into execution only as a DELIBERATE transition after the plan + cold-start prompt exist, never an implicit slide.
-- **Cost lanes.** Route every step to the **cheapest adequate executor** — L0 deterministic script (the batched gate matrix over `gates.json`, the rotation `--check`s) · L1 cheap subagent (extraction/drafting only; the orchestrator verifies) · L2 subscription bridge · L3 frontier judgment. A step with **no named guardrail does not move down** a lane, and the **red lines never move down** (council review models · real code · ADR/handover/changelog-entry wording · persuasive copy · go/no-go · the approval asks). Own-error repair: salvage recorded state first (L0/L1, batched), never frontier re-derivation.
+- **Cost lanes.** Route every step to the **cheapest adequate executor** — L0 deterministic script (the batched gate matrix over `gates.json`, the rotation `--check`s) · L1 cheap subagent (extraction/drafting only; the orchestrator verifies) · L2 subscription bridge · L3 frontier judgment. A step with **no named guardrail does not move down** a lane, and the **red lines never move down** (council review models · real code · ADR/handover/changelog-entry wording · persuasive copy · go/no-go · the approval asks). Own-error repair: salvage recorded state first (L0/L1, batched), never frontier re-derivation. **Prompt economy:** read-only fan-out (research/sweeps/extraction) runs ONLY on restricted-tool vehicles — a full-tool subagent for read-only work is a forbidden lane downgrade (invisible prompt-flood + blast radius), and a subagent is never told to shell out for facts obtainable read-only; the orchestrator's own shell form is ONE plain pipeline per call (a `;`/`&&` chain or env-prefixed invocation never matches a prefix allow rule); a fan-out launcher that gates per call yields to the agent-spawn lane — capability-gated: without restricted-tool vehicles (generic full-tool spawning does not count), read-only research stays in the orchestrator's own context, never a vehicle mandate a host cannot satisfy. Judgment, code, synthesis stay at the frontier lane (a task that genuinely runs/writes keeps a full-tool subagent); honest limit: no deterministic gate classifies a dispatch — canon at the point of use + placed vehicles + the retro loop.
 
 ---
 

package/references/templates/autonomy.json

@@ -0,0 +1,3 @@
+{
+  "_README": "Per-project autonomy policy: red-lines (always) + per-activity autonomy level. Hand-editable; or use the set-autonomy writer (previews, then writes valid JSON). Strict JSON — no comments."
+}