FE-811: Ship-gate rename residue and live mention autocomplete

.fixtures/runs/fixture-curation/fixture-curation-2026-06-05T104440Z/report.json

@@ -72,6 +72,6 @@
     "sessionJsonl": "runs/fixture-curation/fixture-curation-2026-06-05T104440Z/session.jsonl",
     "transcriptMarkdown": "runs/fixture-curation/fixture-curation-2026-06-05T104440Z/transcript.md",
     "reportJson": "runs/fixture-curation/fixture-curation-2026-06-05T104440Z/report.json",
-    "graphSnapshotJson": "runs/fixture-curation/fixture-curation-2026-06-05T104440Z/graph-snapshot.json"
+    "graphOverviewJson": "runs/fixture-curation/fixture-curation-2026-06-05T104440Z/graph-overview.json"
   }
 }

.fixtures/runs/project-graph-review-cycle/2026-06-06-project-graph-review-cycle/graph-snapshot.json → .fixtures/runs/project-graph-review-cycle/2026-06-06-project-graph-review-cycle/graph-overview.json

@@ -1075,4 +1075,3 @@
   "edgeCount": 10,
   "lsn": 4
 }
-

.fixtures/runs/project-graph-review-cycle/2026-06-06-project-graph-review-cycle/report.json

@@ -85,7 +85,7 @@
   ],
   "productUpdates": [
     {
-      "topic": "workspace.snapshot",
+      "topic": "workspace.state",
       "specId": 1,
       "sessionId": "019e9d3e-7f21-76ae-bbc2-2955f779cdac"
     },
@@ -121,6 +121,6 @@
     "sessionJsonl": "runs/project-graph-review-cycle/2026-06-06-project-graph-review-cycle/session.jsonl",
     "transcriptMarkdown": "runs/project-graph-review-cycle/2026-06-06-project-graph-review-cycle/transcript.md",
     "reportJson": "runs/project-graph-review-cycle/2026-06-06-project-graph-review-cycle/report.json",
-    "graphSnapshotJson": "runs/project-graph-review-cycle/2026-06-06-project-graph-review-cycle/graph-snapshot.json"
+    "graphOverviewJson": "runs/project-graph-review-cycle/2026-06-06-project-graph-review-cycle/graph-overview.json"
   }
 }

.fixtures/seeds/bilal-port/duplicate-edge-policy.ts

@@ -14,7 +14,7 @@
  * counted so the porter can keep duplicate-drop stats visible.
  */
 
-export type SeedPortEdgeOrigin = 'source' | 'synthetic';
+ type SeedPortEdgeOrigin = 'source' | 'synthetic';
 
 /** The fields that identify an edge for duplicate detection. */
 export interface SeedEdgeIdentity {

.fixtures/seeds/brunch-self/README.md

@@ -0,0 +1,30 @@
+# `.fixtures/seeds/brunch-self/`
+
+A **faithful** spec graph hand-derived from this repository's own planning prose
+(`memory/SPEC.md` + `memory/PLAN.md`), as opposed to the synthetic
+coverage/edge-spread fixtures.
+
+Purpose:
+
+- prove the whole loop end-to-end: real prose → graph fixture → the real
+  propose-graph validator (`seedFixture` → `CommandExecutor`) → renderers
+- give the renderers a realistic, all-planes anchor to project from, with
+  meaningful titles and rationales instead of synthetic placeholders
+- serve as the worked example / template for porting other projects' spec/plan
+  docs into structurally-legal seed graphs
+
+Coverage (a by-product of being faithful, not the goal):
+
+- every node kind across all four planes (intent / oracle / design / plan)
+- every edge category (dependency, proof, support, realization, boundary,
+  composition, association, supersession), including both proof/support stances
+- one supersession lineage (per-strategy offer-first supersedes the retired
+  universal per-turn ritual)
+
+Contents:
+
+- `spec-graph.json` — one `planning_ready` spec describing Brunch itself.
+
+Structural legality is enforced by the seed loader: `spec-graph` is committed
+through `CommandExecutor` by `src/renderers/graph/previews.test.ts`, which fails
+if any node/edge is structurally illegal.

.fixtures/seeds/dumpchat/README.md

@@ -0,0 +1,46 @@
+# `.fixtures/seeds/dumpchat/`
+
+A spec graph hand-derived from the **dumpchat** project
+(`/Users/lunelson/Code/lunelson/dumpchat`), a WXT browser extension that exports
+ChatGPT / Claude / Perplexity conversations to Markdown via each platform's
+native per-turn copy buttons.
+
+Faithful vs. projected:
+
+- **intent** plane — substantially **faithful** to `docs/SPEC.md` and `README.md`:
+  the copy-button thesis, the four-step extraction flow, index alternation,
+  depth filtering, selector-stability constraints, and the Verify Export
+  diagnostics requirement are all drawn from real prose.
+- **design** plane — **faithful**: nodes map to actual modules
+  (`dumpchat.content.ts`, `lib/dumpchat/extraction.ts`, `config.ts`, `sites/*`)
+  and the `SiteConfig` type contract.
+- **oracle** plane — **mixed**: the `extraction.test.ts` check and the in-page
+  Verify Export run / diagnostics JSON are real; the per-platform re-verification
+  obligation is **projected**.
+- **plan** plane — **substantially projected**: the source has no plan doc, so
+  milestone / frontier / slice nodes are plausible projections from the intent,
+  marked `source: "projected"`.
+
+`readiness_grade` is `commitments_ready`: the source spec commits firmly to
+decisions, invariants, and selector policy, but carries no explicit plan.
+
+Coverage (a by-product of being faithful, not the goal):
+
+- all four planes (intent / oracle / design / plan) and every node kind used in
+  the intent plane
+- every edge category (dependency, realization, boundary, composition,
+  association, supersession, proof, support), including both proof/support
+  stances
+- one supersession lineage: depth-based separation and the modal-depth filter
+  decision supersede the retired `button.closest("pre, code")` check
+
+Contents:
+
+- `spec-graph.json` — 41 nodes / 33 edges (40 / 31 in active context after the
+  superseded predecessor is hidden).
+
+Validate with:
+
+```
+npx tsx src/graph/validate-fixture.ts dumpchat/spec-graph
+```