@event4u/agent-config 14.22.0 → 15.0.0

package/dist/agent-src/skills/token-optimizer/SKILL.md

@@ -78,7 +78,7 @@ INTENT
 │
 └── API-bill lever (stable prefix reused across calls · non-interactive
     bulk cohort · model/effort pick against the paid Anthropic bill)
-      → cite [api-cost-levers](../../../docs/guidelines/agent-infra/api-cost-levers.md)
+      → cite [api-cost-levers](../../guidelines/agent-infra/api-cost-levers.md)
         (prompt caching ~0.1x reads · Batch -50% · model tiering · effort)
 ```
 

package/dist/agent-src/skills/typography-system/SKILL.md

@@ -101,7 +101,7 @@ brand already registers fonts, use them and skip the filter.
 ## Delivery — hosting mode (self-hosted by default)
 
 The delivery policy is **owned** by
-[`design-fidelity-mechanics`](../../../docs/guidelines/design-fidelity-mechanics.md)
+[`design-fidelity-mechanics`](../../guidelines/design-fidelity-mechanics.md)
 § Asset & imagery discipline ([`ADR-205`](../../../docs/decisions/ADR-205-webfont-delivery-ownership.md));
 this skill is a consumer and emits the route, never a competing policy. A font
 CDN link transmits the **visitor's IP** to that third party on every page view —
@@ -174,7 +174,7 @@ the shorter line.
 
 ## See also
 
-- [`design-canon.md`](../../../docs/guidelines/design-canon.md) — named-systems + typography-craft (foundry/theory) grounding index; pull to escape the AI-default fonts.
+- [`design-canon.md`](../../guidelines/design-canon.md) — named-systems + typography-craft (foundry/theory) grounding index; pull to escape the AI-default fonts.
 - [`design-tokens`](../design-tokens/SKILL.md) — toolchain that generates CSS
   vars and the Tailwind snippet from `tokens.json`.
 - [`design-intelligence`](../design-intelligence/SKILL.md) — idiom corpus that

package/dist/agent-src/skills/ui-component-architect/SKILL.md

@@ -46,7 +46,7 @@ Do NOT use when:
 
 Extract a **stateful component** only when an element **repeats ~4× or more AND
 carries real props/state** (the stateful-component bar of the per-class canon,
-[`abstraction-thresholds`](../../../docs/guidelines/abstraction-thresholds.md) —
+[`abstraction-thresholds`](../../guidelines/abstraction-thresholds.md) —
 it does not lower the code-level two-repetition bar). One or two repeats, or a
 repeat with no varying props, is not a component — inline it (a props-only
 shell earns extraction at 3+ uses). A long single-file body is **normal**, not a smell;
@@ -69,13 +69,33 @@ still exists in the tree. Absent or empty → fall through to
 Then review existing components in the codebase for the same
 responsibility — extend rather than rebuild when a match is found.
 
-**A new component's workshop is part of its design, on every React lane and not
-only `react-shadcn`.** When the extraction is agreed, the story set comes from
-[`storybook-workshop`](../storybook-workshop/SKILL.md) — one story per concept
-off the state-coverage matrix below, which is what makes the component
-agent-readable and its contrast floor checkable before it reaches a page. That
-skill is stack-agnostic and ships in `engineering-base`, so a plain-`react`
-project receives it exactly as a `react-shadcn` one does.
+**A new component's workshop is part of its design where the project has one —
+size-conditional, never a generic mandate.** The boundary is
+[`component-oriented-and-oop-development`](../../guidelines/component-oriented-and-oop-development.md)
+§ Component workshop: a workshop pays for itself on a real, growing shared
+component library and is not worth the setup for a handful of one-off
+components. Where it applies, the story set comes from
+[`storybook-workshop`](../storybook-workshop/SKILL.md) — one story per concept,
+which is what makes the component agent-readable and its contrast floor
+checkable before it reaches a page. That skill is stack-agnostic and ships in
+`engineering-base`, so a plain-`react` project receives it exactly as a
+`react-shadcn` one does. The state-coverage matrix the story set is derived
+from is React-specific and lives in
+[`react-shadcn-ui`](../react-shadcn-ui/SKILL.md) § Step 3, deliberately not
+duplicated here; a Blade, Vue or Livewire lane derives its story set from that
+lane's own state list instead.
+
+**No order is prescribed among stories, component and page — and that is a
+positive statement, not an omission.** The observed pattern, described rather
+than required: when porting existing UI the first move is inventory, not
+construction (this lifecycle excludes Plan by definition), and stories are
+written concurrent with the component rather than as a phase gate before or
+after it. No primary source states a required order, Storybook's own tutorial
+writes the component before the story, Atomic Design's author calls it "not a
+linear process", and measured evidence comparing build orders is **absent**,
+not merely weak. A claimed order is therefore a preference someone holds, never
+this skill's instruction.
+
 If none exists, write the new component's purpose: *"Renders a
 labelled input with inline error and hint."* If the sentence has an
 "and" joining two unrelated jobs, the component is two components.

package/dist/agent-src/skills/unit-economics-modeling/SKILL.md

@@ -34,7 +34,7 @@ Do NOT use for full-business intrinsic-value modeling, OKR setting, funnel-stage
   customer and one fully-loaded acquisition dollar. Aggregate ratios
   ride on per-unit truth; if the unit is mis-defined (trial vs paid,
   household vs seat), every ratio downstream is decoration. See
-  [`docs/contracts/mental-models.md`](../../../docs/contracts/mental-models.md) § 1.
+  [`docs/contracts/mental-models.md`](../../docs/contracts/mental-models.md) § 1.
 - **Mental model 8 — Second-order thinking.** A CAC drop driven by
   discounting lifts LTV/CAC on paper while shortening cohort
   retention — the second-order effect lands two quarters later in
@@ -46,7 +46,7 @@ Do NOT use for full-business intrinsic-value modeling, OKR setting, funnel-stage
   **fiscal-period** slot for the close-window the ratios must
   reconcile against (monthly close vs quarterly board pack vs
   annual plan). See
-  [`context-spine`](../../../docs/contracts/context-spine.md).
+  [`context-spine`](../../docs/contracts/context-spine.md).
 
 ## Procedure
 

package/dist/agent-src/skills/upstream-contribute/SKILL.md

@@ -222,7 +222,7 @@ If not in the package repo, note that these checks will run in CI after the PR i
 
 If the contribution originates from a proposal under `agents/proposals/`
 (produced by `learning-to-rule-or-skill` via the pipeline in
-[`self-improvement-pipeline`](../../../docs/guidelines/agent-infra/self-improvement-pipeline.md)),
+[`self-improvement-pipeline`](../../guidelines/agent-infra/self-improvement-pipeline.md)),
 run the Stage-4 gate before opening the PR:
 
 ```bash

package/dist/agent-src/skills/validate-feature-fit/SKILL.md

@@ -53,7 +53,7 @@ Verify the request doesn't conflict with:
 - Data flow (does it bypass existing services or repositories?)
 - **Product rules and domain invariants** — pull active rules via the
   shared abstraction (see
-  [`memory-access`](../../../docs/guidelines/agent-infra/memory-access.md)):
+  [`memory-access`](../../guidelines/agent-infra/memory-access.md)):
 
   ```bash
   agent-config memory:lookup \
@@ -66,7 +66,7 @@ Verify the request doesn't conflict with:
   (e.g., "free plan caps at 3 users"); the feature must either respect
   it or explicitly propose to retire it. Cite the matching `id:` in
   the findings. Schema:
-  [`engineering-memory-data-format`](../../../docs/guidelines/agent-infra/engineering-memory-data-format.md).
+  [`engineering-memory-data-format`](../../guidelines/agent-infra/engineering-memory-data-format.md).
 
 **If contradiction found** → show the existing pattern, explain why it matters.
 

package/dist/agent-src/skills/verify-repair-loop/SKILL.md

@@ -156,7 +156,7 @@ iterating against a stale contract.
 
 State lives in the conversation, optionally mirrored to a re-read working file —
 **never** a daemon or persistent cross-session store (honors
-[`resident-process-governance`](../../../docs/contracts/resident-process-governance.md)).
+[`resident-process-governance`](../../docs/contracts/resident-process-governance.md)).
 
 ## Validation
 
@@ -261,7 +261,7 @@ iter2 fixes A, breaks B, 92% → "PASS"      ✗ regression ignored
   `tests/fixtures/tamper-corpus/` and its polarity proven both ways by
   `--self-test`.
 * [`autonomous-execution`](../../rules/autonomous-execution.md) — the N=3 cap.
-* [`resident-process-governance`](../../../docs/contracts/resident-process-governance.md) — the
+* [`resident-process-governance`](../../docs/contracts/resident-process-governance.md) — the
   runtime-free constraint this loop honors.
 * Optional impact pre-step (when a code-graph is present, ADR-124): run
   `agent-config code-graph affected --since <ref>` to scope which symbols a

package/dist/agent-src/skills/vision-articulation/SKILL.md

@@ -26,7 +26,7 @@ Do NOT use for outward-facing fundraise pitch (route to Wing-3 `fundraising-narr
 
 ## Cognition cluster
 
-- **Mental model 27 — Why now.** Vision without a *why now* is a wishlist. The market shift, technology wave, regulatory change, or demographic inflection that makes the vision *achievable in this window* is the load-bearing claim. See [`mental-models.md`](../../../docs/contracts/mental-models.md) § 27.
+- **Mental model 27 — Why now.** Vision without a *why now* is a wishlist. The market shift, technology wave, regulatory change, or demographic inflection that makes the vision *achievable in this window* is the load-bearing claim. See [`mental-models.md`](../../docs/contracts/mental-models.md) § 27.
 - **Mental model 1 — First principles.** Strip the vision to outcomes for a named cohort, not feature lists or company milestones. *"Be the X for Y"* is shape; *"customer C does outcome O 10× faster"* is substance. See `mental-models.md` § 1.
 - **Mental model 21 — Second-order thinking.** *"If this vision is realised, what does the world look like in 5 years?"* If the answer is just *"we got bigger"*, the vision isn't load-bearing; if the answer is structural change (workflow shift, market re-shape), it is. See `mental-models.md` § 21.
 - **Context-spine — org-stage + product + customer-segment.** Read **org-stage** for vision-horizon (pre-seed = 3-year survival vision; growth = 5–7-year market-shape vision; mature = 10-year). Read **product** for the realistic shipping shape. Read **customer-segment** for the cohort whose outcome anchors the vision.

package/dist/agent-src/skills/voc-extract/SKILL.md

@@ -47,7 +47,7 @@ quantitative funnel diagnosis ([`funnel-analysis`](../funnel-analysis/SKILL.md))
 - **Mental model 15 — Signal vs noise.** A loud single reporter
   swamps quiet recurring patterns; rank by **distinct authors**, not
   comment count. See
-  [`docs/contracts/mental-models.md`](../../../docs/contracts/mental-models.md) § 15.
+  [`docs/contracts/mental-models.md`](../../docs/contracts/mental-models.md) § 15.
 - **Mental model 14 — Pareto.** Roughly 20% of themes carry 80% of
   the contact volume; cut the long tail explicitly so the team acts
   on the head. See `mental-models.md` § 14.
@@ -57,7 +57,7 @@ quantitative funnel diagnosis ([`funnel-analysis`](../funnel-analysis/SKILL.md))
 - **Product context-spine slot.** Read **product** for segments,
   non-goals, and focal jobs; do not surface themes that fall outside
   the declared scope without a scope-violation flag. See
-  [`context-spine`](../../../docs/contracts/context-spine.md).
+  [`context-spine`](../../docs/contracts/context-spine.md).
 
 ## Procedure
 

package/dist/agent-src/skills/voice-and-tone-design/SKILL.md

@@ -36,7 +36,7 @@ subject lines or ad creative formats (channel-agnostic skill).
   segment can recognise across surfaces is the voice that survives
   contact with three writers under deadline. A five-attribute voice
   that no one remembers is brand-theatre. See
-  [`docs/contracts/mental-models.md`](../../../docs/contracts/mental-models.md) § 13.
+  [`docs/contracts/mental-models.md`](../../docs/contracts/mental-models.md) § 13.
 - **Mental model 15 — Signal vs. noise.** Voice is the
   identification signal embedded in copy — without distinctiveness
   it is noise that competes with every peer's noise. Cut to the
@@ -46,7 +46,7 @@ subject lines or ad creative formats (channel-agnostic skill).
   for the proofs the voice must carry (a voice that promises what
   the product cannot back is fiction); read **customer-segment**
   for the listening-register the audience already lives in. See
-  [`context-spine`](../../../docs/contracts/context-spine.md).
+  [`context-spine`](../../docs/contracts/context-spine.md).
 
 ## Procedure
 

package/dist/agent-src/skills/wireframe/SKILL.md

@@ -101,7 +101,7 @@ branded, recommendation-bearing). A single refined implementation →
 - Do NOT skip decision capture — an exploration round without a recorded
   decision has to be re-litigated in the hi-fi round.
 - Do NOT embed system internals (paths, skill/tool names) in the artifact —
-  Q13 in [`design-antipatterns.md`](../../../docs/guidelines/design-antipatterns.md).
+  Q13 in [`design-antipatterns.md`](../../guidelines/design-antipatterns.md).
 
 ## Gotcha
 

package/dist/agent-src/skills/workspace-link/SKILL.md

@@ -127,5 +127,5 @@ A workspace-link answer MUST contain, in order:
 ## See also
 
 - [`monorepo-workspace`](../monorepo-workspace/SKILL.md) — which manager, which workspaces, which graph.
-- [`monorepo-antipatterns`](../../../docs/guidelines/monorepo-antipatterns.md) — the task-graph mistakes a `paths` patch hides.
+- [`monorepo-antipatterns`](../../guidelines/monorepo-antipatterns.md) — the task-graph mistakes a `paths` patch hides.
 - [`blast-radius-analyzer`](../blast-radius-analyzer/SKILL.md) — § Monorepo, once the edge exists.

package/dist/agent-src/templates/agent-settings.md

@@ -14,7 +14,7 @@ Keep the format regular — 2-space indent, no tabs, no lists, one nesting
 level only. The installer's YAML handler is a restricted stdlib parser, not
 a full YAML engine. Ask the agent to normalize after manual edits — it
 follows the merge rules in
-[`layered-settings`](../docs/guidelines/agent-infra/layered-settings.md#section-aware-merge-rules).
+[`layered-settings`](../guidelines/agent-infra/layered-settings.md#section-aware-merge-rules).
 
 ### Migration from the legacy `.agent-settings` (key=value)
 
@@ -36,7 +36,7 @@ This block defines the personal and project-level settings that
 `scripts/install.ts` (via `config/agent-settings.template.yml`)
 writes to `.agent-settings.yml` on first install. Subsequent edits are
 made by the user directly or by the agent on request, following the
-[section-aware merge rules](../docs/guidelines/agent-infra/layered-settings.md#section-aware-merge-rules).
+[section-aware merge rules](../guidelines/agent-infra/layered-settings.md#section-aware-merge-rules).
 
 ```yaml
 # Agent Settings
@@ -597,7 +597,7 @@ linked_projects_max_files: 20000
 
 Personal and project-level settings (initial file written by
 `scripts/install.ts`, edits follow the merge rules in
-[`layered-settings`](../docs/guidelines/agent-infra/layered-settings.md#section-aware-merge-rules)).
+[`layered-settings`](../guidelines/agent-infra/layered-settings.md#section-aware-merge-rules)).
 **Key paths use dot-notation** to denote nesting: `personal.user_name`
 lives under `personal:` in YAML.
 
@@ -636,9 +636,9 @@ the canonical narrative lives in
 | `subagents.implementer_model` | model alias or empty | _(empty)_ | Model for implementer subagents. Empty = same tier as session model. See [subagent-configuration](../contexts/subagent-configuration.md). |
 | `subagents.judge_model` | model alias or empty | _(empty)_ | Model for judge subagents. Empty = one tier above implementer (opus if sonnet, sonnet if haiku). |
 | `subagents.max_parallel` | integer | `3` | Maximum parallel subagent invocations. `1` serializes. |
-| `roles.default_role` | `""`, `developer`, `reviewer`, `tester`, `po`, `incident`, `planner` | _(empty)_ | Role the agent defaults to at the start of a session. Lives in the PROJECT settings (`agent-project-settings.example.yml` § roles), not in the personal template. See [`role-contracts`](../docs/guidelines/agent-infra/role-contracts.md). |
+| `roles.default_role` | `""`, `developer`, `reviewer`, `tester`, `po`, `incident`, `planner` | _(empty)_ | Role the agent defaults to at the start of a session. Lives in the PROJECT settings (`agent-project-settings.example.yml` § roles), not in the personal template. See [`role-contracts`](../guidelines/agent-infra/role-contracts.md). |
 | `roles.active_role` | same as `default_role` | _(empty)_ | Role currently active; set by `/mode <name>`, cleared by `/mode none`. Enables the `role-mode-adherence` rule. |
-| `personas.override` | list of persona ids | `[]` | Developer-local override of the team default lens cast (not in the shipped personal template — add the block when needed). Empty = inherit `personas.default` from `.agent-project-settings.yml`. See [`layered-settings`](../docs/guidelines/agent-infra/layered-settings.md). |
+| `personas.override` | list of persona ids | `[]` | Developer-local override of the team default lens cast (not in the shipped personal template — add the block when needed). Empty = inherit `personas.default` from `.agent-project-settings.yml`. See [`layered-settings`](../guidelines/agent-infra/layered-settings.md). |
 | `personas.ignore` | list of persona ids | `[]` | Persona ids dropped from the default cast locally. Ignored personas stay invokable via `--personas=<id>`. |
 | `onboarding.onboarded` | `true`, `false` | `false` | Whether `/onboard` has run on this project. The `onboarding-gate` rule prompts for `/onboard` when this is `false`. Missing entirely = legacy project, treated as onboarded. |
 | `commands.suggestion.enabled` | `true`, `false` | `true` | Master switch for the command-suggestion layer. `false` = the layer is silent; explicit `/commands` still work. See `rules/command-suggestion-policy.md`. |
@@ -732,7 +732,7 @@ they ship with a live consumer in code and get documented here, not before.
 ## Sync rules
 
 When new settings are added to this template, the
-[section-aware merge rules](../docs/guidelines/agent-infra/layered-settings.md#section-aware-merge-rules)
+[section-aware merge rules](../guidelines/agent-infra/layered-settings.md#section-aware-merge-rules)
 govern the update:
 
 1. Missing keys are added with their **default value** from this template,

package/dist/agent-src/templates/agents/agent-project-settings.example.yml

@@ -39,7 +39,7 @@ schema_version: 1
 # CI guard: a release bump of `package.json` must update this value
 # in lockstep — see scripts/check_template_pin_drift.ts (road-to-
 # portable-runtime-and-update-check P3.3).
-agent_config_version: "14.22.0"
+agent_config_version: "15.0.0"
 
 # --- Project identity ---
 project:

package/dist/agent-src/templates/contexts/knowledge-pages.md

@@ -91,9 +91,8 @@ tiny pages do.
 
 ## What this substrate is — honest framing
 
-A **governed second-brain substrate**: working-memory continuity across
-compactions and sessions (hot-context cache, dispatcher-wide on all 7
-platforms), deterministic where safe (dedup, recurrence, index regen,
+A **governed second-brain substrate**: a memory session index restored at
+`session_start` across all 7 platforms, deterministic where safe (dedup, recurrence, index regen,
 fold rollups, scale tripwires), human-gated where meaning is at stake
 (consolidate gate, promotion, contested flow). It is **not**
 self-organizing intelligence — the substrate is governed files, and the

package/dist/agent-src/templates/copilot-instructions.md

@@ -111,7 +111,7 @@ none. **Intentional, not a bug.** The installer projects
 resolve at agent runtime. The patterns below are correct by design:
 
 - **Relative cross-references inside markdown rules / skills /
-  commands** like `[x](../docs/guidelines/foo.md)` or
+  commands** like `[x](../guidelines/foo.md)` or
   `[y](../contexts/bar.md)`. The base resolves from the file's
   delivered location (e.g. `.augment/rules/<rule>.md`) — these paths
   are valid via the `.augment/` tree, even when the file is symlinked

package/dist/agent-src/templates/roadmaps.md

@@ -524,7 +524,7 @@ that was never a judgement call.
     No material risks → the **exact** honest-null grammar; a bare "no
     risks" sentence fails. Schema, staleness rule, and the grandfather
     clause for pre-gate roadmaps:
-    [`plan-review-gates § 1`](../../docs/contracts/plan-review-gates.md).
+    [`plan-review-gates § 1`](../docs/contracts/plan-review-gates.md).
     `status: draft` roadmaps are exempt until flipped to ready. Enforced
     by `lint_plan_risk_register` at pre-push + CI.
 

package/dist/agent-src/templates/scripts/README.md

@@ -111,7 +111,7 @@ _Data source: ownership-map.yml + historical-bug-patterns.yml. …_
 ```
 
 Curate `ownership-map.yml` and `historical-bug-patterns.yml` over time —
-see `../../../docs/guidelines/agent-infra/review-routing-data-format.md` for
+see `../../guidelines/agent-infra/review-routing-data-format.md` for
 the full schema.
 
 ## Uninstall
@@ -191,5 +191,5 @@ whatever is under `agents/memory/`. Trigger manually with
 Exit `0` = clean, `1` = violations (missing required fields, duplicate
 ids, or obvious secrets — staleness alone is informational).
 
-Schema reference: `../../../docs/guidelines/agent-infra/engineering-memory-data-format.md`.
+Schema reference: `../../guidelines/agent-infra/engineering-memory-data-format.md`.
 Schema examples: `.augment/templates/agents/memory/*.example.yml`.

package/dist/agent-src/user-types/_template/user-type.md

@@ -87,7 +87,7 @@ Non-negotiable. Guardrails are encoded here.
 
 *Author note (delete before publishing): this template targets the
 7-section spine locked in
-[`docs/contracts/user-type-schema.md`](../../../docs/contracts/user-type-schema.md).
+[`docs/contracts/user-type-schema.md`](../../docs/contracts/user-type-schema.md).
 Stay within the **≤ 120 line** budget (file total, including
 frontmatter). Replace every `{placeholder}` with concrete content
 that passes the Anti-Generic Quality Bar (≥ 5 concrete review

package/dist/cli/registry.js

@@ -73,7 +73,7 @@ export const REGISTRY = [
     { name: 'affected', disposition: 'delegate', synopsis: 'Artefacts related to <artefact> via the discovery relation-graph (BFS).' },
     { name: 'graph-explain', disposition: 'delegate', synopsis: 'Seed on a <concept>, expand 2 hops over the discovery relation-graph with a node budget.' },
     { name: 'benchmark', disposition: 'delegate', synopsis: 'Report context-token reduction vs the full always-loaded projection.' },
-    { name: 'code-graph', disposition: 'delegate', synopsis: 'Deterministic code-graph engine: build|detect|query|affected|path|explain|validate.' },
+    { name: 'code-graph', disposition: 'delegate', synopsis: 'Deterministic code-graph engine: build|detect|refresh|validate|query|affected|path|explain, plus the gate verbs impact|tests-for|untested|dead.' },
     { name: 'hooks:install', disposition: 'delegate', synopsis: 'Install the combined pre-commit hook.' },
     { name: 'hooks:status', disposition: 'delegate', synopsis: 'Print the runtime hook matrix (per-platform install + bindings).' },
     { name: 'hooks:doctor', disposition: 'delegate', synopsis: 'Diagnose hook health: concerns, posture, missing trampolines (read-only).' },

package/dist/cli-delegate/archive_completed_roadmaps.js

@@ -1,7 +1,9 @@
 #!/usr/bin/env tsx
 import { createRequire as __acCreateRequire } from 'node:module'; const require = globalThis.require ?? __acCreateRequire(import.meta.url);
 import {
-  collect
+  blocker_is_resolved,
+  collect,
+  parse_blockers
 } from "./chunk-QAJS5YWY.js";
 import "./chunk-ULET4BYS.js";
 import {
@@ -132,6 +134,38 @@ function _git_mv(root, src_rel, dst_rel, dry_run) {
     return { ok: false, tracked: false };
   }
 }
+function _blockerStatuses(text) {
+  const out = /* @__PURE__ */ new Map();
+  for (const b of parse_blockers(text)) {
+    out.set(b.id, blocker_is_resolved(b) ? "resolved" : "open");
+  }
+  return out;
+}
+function _indexContent(root, rel) {
+  const cp = _run(["git", "show", `:${rel}`], root);
+  return cp.returncode === 0 ? cp.stdout : null;
+}
+function _blockerDivergence(root, rel, workTreeText) {
+  const staged = _indexContent(root, rel);
+  if (staged === null) {
+    return [];
+  }
+  const wt = _blockerStatuses(workTreeText);
+  const idx = _blockerStatuses(staged);
+  const ids = [.../* @__PURE__ */ new Set([...wt.keys(), ...idx.keys()])].sort();
+  const out = [];
+  for (const id of ids) {
+    const a = idx.get(id);
+    const b = wt.get(id);
+    if (a === b) {
+      continue;
+    }
+    out.push(
+      `blocker '${id}': staged=${a ?? "absent"} vs working tree=${b ?? "absent"}`
+    );
+  }
+  return out;
+}
 function _fs_grep_files(root, needle) {
   const SKIP = /* @__PURE__ */ new Set([".git", "node_modules", ".augment", "dist"]);
   const hits = [];
@@ -275,9 +309,8 @@ function archive_completed(root, opts) {
   );
   const archived = [];
   for (const stats of collect(roadmap_root)) {
-    const guarded = parseGuardedBaselines(
-      fs.readFileSync(path.join(roadmap_root, stats.rel), "utf-8")
-    );
+    const roadmapText = fs.readFileSync(path.join(roadmap_root, stats.rel), "utf-8");
+    const guarded = parseGuardedBaselines(roadmapText);
     if (guarded.length > 0) {
       process.stderr.write(
         `  \u26A0\uFE0F  ${stats.rel}: ${guarded.length} guarded-baseline step(s) \u2014 not archived. Verify against the real mechanism, then flip to \`[x]\`.
@@ -320,6 +353,18 @@ function archive_completed(root, opts) {
     if (changed_only && !touched.has(old_rel)) {
       continue;
     }
+    const divergence = _blockerDivergence(root, old_rel, roadmapText);
+    if (divergence.length > 0) {
+      process.stderr.write(
+        `  \u26A0\uFE0F  ${stats.rel}: staged and working-tree blockers disagree (${divergence.length}) \u2014 not archived. \`git mv\` would commit the staged content, which this sweep never validated. Stage or discard the change, then re-run.
+`
+      );
+      for (const d of divergence) {
+        process.stderr.write(`        \xB7 ${d}
+`);
+      }
+      continue;
+    }
     const new_rel = `agents/roadmaps/archive/${stats.rel}`;
     const mv = _git_mv(root, old_rel, new_rel, dry_run);
     if (!mv.ok) {
@@ -327,6 +372,9 @@ function archive_completed(root, opts) {
 `);
       continue;
     }
+    if (!dry_run && mv.tracked) {
+      _run(["git", "add", "--", new_rel], root);
+    }
     const refs = _inbound_ref_rewrite(root, old_rel, new_rel, dry_run, mv.tracked);
     if (!dry_run && refs.length && mv.tracked) {
       _run(["git", "add", "--", ...refs], root);

package/dist/cli-delegate/check_update_banner.js

@@ -218,7 +218,18 @@ var MERGEABLE_KEYS = [
   "knowledge.global_sharing.redaction.halt_on_trigger",
   "knowledge.global_sharing.auto_promote_threshold",
   "knowledge.global_sharing.freshness.hypothesis_after_days",
-  "knowledge.global_sharing.freshness.stale_after_days"
+  "knowledge.global_sharing.freshness.stale_after_days",
+  // `design.fidelity_mode` is a per-DEVELOPER working preference, not a
+  // per-project one: whether an agent may deviate from a handed-over design
+  // is a property of how that person works, and a designer who sets it once
+  // expects it in every checkout. `design-fidelity.md` already tells the
+  // reader to resolve it through the cascade — "that file is the project
+  // layer of a cascade that starts user-global" — and the cascade then
+  // filtered it out here, silently, so the rule's own instruction could not
+  // be followed. Class C, so a user still sets it by hand or through the
+  // GUI; whitelisting decides only whether the value SURVIVES the merge
+  // (ADR-271, road-to-design-intent-conformance 2.3).
+  "design.fidelity_mode"
 ];
 var TEMPLATE_RELATIVE = path2.join("src", "config", "agent-settings.template.yml");
 var _PACKAGE_ROOT = path2.resolve(path2.dirname(fileURLToPath(import.meta.url)), "..", "..", "..");

package/dist/cli-delegate/{chunk-TGIPZBCW.js → chunk-2KKXJW25.js}

@@ -341,22 +341,35 @@ function listContinuityRecords(workspace_root) {
 }
 function resolveContinuityRecord(workspace_root, session_id) {
   const id = String(session_id ?? "").trim();
-  if (id !== "") {
-    const own = path3.join(workspace_root, recycle_envelope_rel(id));
-    return fs3.existsSync(own) ? { file: own, reason: `continuity record for this session (${safe_stem(id)})` } : { file: null, reason: `no continuity record for this session (${safe_stem(id)})` };
-  }
-  const candidates = listContinuityRecords(workspace_root);
+  const own = id === "" ? null : path3.join(workspace_root, recycle_envelope_rel(id));
+  const candidates = listContinuityRecords(workspace_root).filter((f) => f !== own);
   if (candidates.length === 0) {
-    return { file: null, reason: "no continuity record present" };
+    return {
+      file: null,
+      reason: id === "" ? "no continuity record present" : `no predecessor record in this workspace (own: ${safe_stem(id)})`
+    };
   }
   if (candidates.length === 1) {
-    return { file: candidates[0], reason: "one continuity record, no session id to key on" };
+    return {
+      file: candidates[0],
+      reason: id === "" ? "one continuity record, no session id to key on" : `predecessor record, one candidate (own: ${safe_stem(id)})`
+    };
   }
   return {
     file: null,
-    reason: `starting clean: ${candidates.length} continuity records in this workspace and no session id to tell them apart \u2014 resuming from the newest would resume a peer session, not this one`
+    reason: `starting clean: ${candidates.length} continuity records in this workspace and no way to tell which is the predecessor \u2014 resuming from the newest would resume a peer session, not this one`
   };
 }
+function resolvePredecessor(projectRoot, sessionId) {
+  try {
+    const consumed = path3.join(projectRoot, recycle_consumed_rel(sessionId));
+    const raw = JSON.parse(fs3.readFileSync(consumed, "utf-8"));
+    const prior = String(raw.session_id ?? "").trim();
+    return prior === "" ? "none" : prior;
+  } catch {
+    return "none";
+  }
+}
 function predecessorTracePresent(workspace_root, predecessor) {
   const id = String(predecessor ?? "").trim();
   if (id === "" || id === "none") {
@@ -1104,6 +1117,7 @@ export {
   recycle_envelope_rel,
   recycle_consumed_rel,
   resolveContinuityRecord,
+  resolvePredecessor,
   predecessorTracePresent,
   env_session_id
 };

package/dist/cli-delegate/{chunk-NLP54U3J.js → chunk-2XY7ILBT.js}

@@ -170,7 +170,18 @@ var MERGEABLE_KEYS = [
   "knowledge.global_sharing.redaction.halt_on_trigger",
   "knowledge.global_sharing.auto_promote_threshold",
   "knowledge.global_sharing.freshness.hypothesis_after_days",
-  "knowledge.global_sharing.freshness.stale_after_days"
+  "knowledge.global_sharing.freshness.stale_after_days",
+  // `design.fidelity_mode` is a per-DEVELOPER working preference, not a
+  // per-project one: whether an agent may deviate from a handed-over design
+  // is a property of how that person works, and a designer who sets it once
+  // expects it in every checkout. `design-fidelity.md` already tells the
+  // reader to resolve it through the cascade — "that file is the project
+  // layer of a cascade that starts user-global" — and the cascade then
+  // filtered it out here, silently, so the rule's own instruction could not
+  // be followed. Class C, so a user still sets it by hand or through the
+  // GUI; whitelisting decides only whether the value SURVIVES the merge
+  // (ADR-271, road-to-design-intent-conformance 2.3).
+  "design.fidelity_mode"
 ];
 var TEMPLATE_RELATIVE = path.join("src", "config", "agent-settings.template.yml");
 var _PACKAGE_ROOT = path.resolve(path.dirname(fileURLToPath(import.meta.url)), "..", "..", "..");
@@ -684,6 +695,7 @@ export {
   LOCAL_PROJECT_SUBDIR,
   project_settings_path,
   USER_GLOBAL_FILENAME,
+  user_global_settings_paths,
   MERGEABLE_KEYS,
   find_project_root_with_anchor,
   find_project_root,

package/dist/cli-delegate/{chunk-4FS4WRXP.js → chunk-3FKR64R5.js}

@@ -4063,7 +4063,8 @@ var confidenceBand = external_exports.enum(["off", "low", "medium", "high"]);
 var onBlock = external_exports.enum(["stop", "ask", "warn"]);
 var onBlockFallback = external_exports.enum(["stop", "warn"]);
 var modelAutoSwitch = external_exports.enum(["auto", "suggest", "off"]);
-var leanProjectionMode = external_exports.enum(["eager-all", "thin"]);
+var leanProjectionMode = external_exports.enum(["eager-all", "thin", "delivery"]);
+var leanProjectionHost = external_exports.enum(["claude-code", "cursor", "cline"]);
 var projectionMode = external_exports.enum(["legacy-all", "scoped"]);
 var memoryCadence = external_exports.enum(["auto", "always", "never"]);
 var projectAudience = external_exports.enum(["self", "internal", "client", "public"]);
@@ -4095,9 +4096,12 @@ var settingsSchema = external_exports.object({
   ),
   lean_projection: external_exports.object({
     mode: leanProjectionMode.default("eager-all").describe(
-      "How the per-tool projector emits the rule layer. eager-all = every rule body inlined into every projection (default, safe). thin = kernel rules full-bodied + non-kernel rules as router-resolved pointers (~45k GPT-tok lighter per session). EXPERIMENTAL: validate with the live A/B before flipping; one-flip revert to eager-all."
+      "How the per-tool projector emits the rule layer. eager-all = every rule body inlined into every projection (safe; what an ABSENT key still resolves to, per ADR-267 decision point 4). thin = kernel rules full-bodied + non-kernel rules as router-resolved pointers (~45k GPT-tok lighter per session), with no delivery mechanism. delivery = thin stubs PLUS the rule-inject hook concern, which delivers a body back on a trigger match; this is the value the SHIPPED TEMPLATE carries for Claude Code since ADR-267, and the activation charge for it is paid in src/config/hook-token-budget.json (user_prompt_submit slot sum 4,096 -> 16,384)."
+    ),
+    hosts: external_exports.array(leanProjectionHost).default(["claude-code"]).describe(
+      "Which hosts a thinning mode may thin (ADR-267). Absent resolves to [claude-code]. Only the three hosts with a per-rule rule tree are accepted, and this enum is the AUTHORING layer: an id outside it is a hard validation FAILURE here and under validate_agent_settings, never a warning. resolveLeanProjectionHosts is the READ-TIME backstop for a hand-edited .agent-settings.yml that passed neither, and THERE the id is dropped with a warning (R2 finding 6 \u2014 the two layers answer differently and neither said so). Every host outside this list receives exactly what eager-all writes, asserted byte-for-byte by check_host_tree_parity."
     )
-  }).default({ mode: "eager-all" }),
+  }).default({ mode: "eager-all", hosts: ["claude-code"] }),
   cost: external_exports.object({
     budgets: external_exports.object({
       daily: external_exports.number().min(0).default(0).describe(
@@ -4193,7 +4197,7 @@ var settingsSchema = external_exports.object({
   }),
   chat_history: external_exports.object({
     enabled: external_exports.boolean().default(true).describe(
-      "Persist a structured log of every chat turn to .agent-config/chat-history/ so /chat-history:show, :import, and :learn can replay sessions. Turn off if you never want chat transcripts on disk."
+      "Persist a structured log of every chat turn to agents/runtime/.agent-chat-history (JSONL). Recording only \u2014 no command in this package replays a session from the file; it is a record on disk that a human or a later tool can read. Turn off if you never want chat transcripts on disk."
     ),
     frequency: chatFreq.default("per_turn").describe(
       "How often the chat-history writer flushes to disk. per_turn = after every user / agent exchange (default, lowest data loss on crash). per_phase = at phase boundaries (cheaper I/O). per_tool = after every tool call (highest fidelity, noisiest log)."
@@ -4259,8 +4263,24 @@ var settingsSchema = external_exports.object({
   design: external_exports.object({
     fidelity_mode: fidelityMode.default("strict").describe(
       "How strictly the agent must follow a user-provided prototype / mockup / design system (consumed by the design-fidelity rule). strict = build 1:1, every visible deviation needs confirmation; structural = structure locked, silent gaps fillable with a stated assumption; hard-floor = any deviation is never autonomous."
-    )
-  }).default({ fidelity_mode: "strict" }),
+    ),
+    approximation: external_exports.object({
+      enabled: external_exports.boolean().default(false).describe(
+        "Whether a value INSIDE the configured tolerance may be replaced by the project's own token without a per-value confirmation. Ships false. The distance from the artifact's value to the nearest project token is computed and REPORTED on every value row regardless of this switch \u2014 what it decides is whether the agent may act on it without asking. Turning it on for a whole install is a consumer-facing default flip and is owner-reserved (agents/roadmaps/stubs/road-to-frontend-power-default-flip.md); design.fidelity_mode: hard-floor overrides it to off."
+      ),
+      tolerance: external_exports.object({
+        color: external_exports.number().nullable().default(null).describe(
+          'Colour distance below which a value may be reconciled, in OKLab \u0394EOK. Ships null, and null means "preserve and report" rather than "reconcile freely". No number is set because none is supported: zero of seven design-to-code benchmarks in the evidence set score token conformance, and a number in a config file \u2014 even flagged unmeasured \u2014 shapes behavior and acquires the authority of a default. Not an RGB channel delta: RGB distance is not perceptually uniform.'
+        ),
+        length: external_exports.number().nullable().default(null).describe(
+          "Length distance below which a value may be reconciled, in absolute CSS pixels. Ships null, same reasoning as color. GitHub Primer's \xB11px spacing widening is recorded as an externally observed candidate in src/scripts/_lib/design_tolerance.ts and is deliberately not the shipped value; max(1px, 2%) was rejected for growing more permissive at large dimensions with no evidence that this is wanted."
+        )
+      }).default({ color: null, length: null })
+    }).default({ enabled: false, tolerance: { color: null, length: null } })
+  }).default({
+    fidelity_mode: "strict",
+    approximation: { enabled: false, tolerance: { color: null, length: null } }
+  }),
   consistency: external_exports.object({
     cross_source: crossSourceMode.default("on").describe(
       "Consumed by the cross-source-consistency rule. When the agent works from multiple sources (ticket text, an attached image/mockup, the spec, the codebase) it checks them against each other and asks before proceeding on a discrepancy \u2014 instead of silently guessing. on (default) = surface every real cross-source contradiction / silent-scope-expansion as one question; auto = surface only high-confidence contradictions, state low-confidence as an assumption; off = no cross-source checking."
@@ -4462,6 +4482,14 @@ var settingsSchema = external_exports.object({
       )
     })
   }),
+  continuity: external_exports.object({
+    auto_record: external_exports.enum(["on", "off"]).default("on").describe(
+      "Deterministic continuity-record writer at session end (road-to-continuity-writer-activation Phase 1). on (default since 2026-09-10) = the session-eol concern writes the continuity_record capsule variant on Stop for a substantive session that has claimed a roadmap; every field is computed from on-disk state, with no model spend and no subprocess. off = no automatic record, and only the explicit agent-config session:recycle writes one. The default flipped once the parity evidence landed, so the normal path carries an automatic writer rather than a remembered command; session:recycle is retained because the automatic writer skips a session that claimed no roadmap and omits the git-status anchors."
+    ),
+    run_checkpoints: external_exports.enum(["on", "off"]).default("on").describe(
+      "Deterministic run-checkpoint production at session end (road-to-continuity-writer-activation Phase 1.4). on (default) = a session above the recycle threshold and inside a running roadmap contract leaves agents/runtime/state/checkpoints/<run>.json, so a killed run resumes from a derived checkpoint rather than from bookkeeping. off = no checkpoint; continuity writing, the recycle advisory and the context-fill surface are unaffected. Default ON because it is the behavior the tree already had \u2014 the switch exists to make the three session-end handlers independently disableable, not to change what ships."
+    )
+  }),
   memory: external_exports.object({
     cadence: memoryCadence.default("always").describe(
       "Cadence of the \u{1F9E0} memory-visibility line after a memory-consulting step. always (default) = show whenever a memory type was asked; auto = show only when 3+ types were consulted (less noise); never = suppress. Distinct from rule_loading_tier \u2014 owns its own key since the 2026-06-01 untangle."