feat(shapes): the practice is written where it is read, and the coverage line measures the slip (milestone 439 step 6)

- reusing-code: "Before the turn ends — say what you built" — the four
  verdicts, the one classify_shapes(repo=…) call, and the component file as a
  shape. Description names the end-of-turn moment.
- shape-accounting: the writer judges; audits are the check that it held.
  The write path SUGGESTS (no more hook instances); component file rows and
  whole-file canon described; scoped covers Svelte too.
- _INSTRUCTIONS reuse line: "before the turn ends, say what you built
  (create_snippet the reusable, classify_shapes the rest)" — 1570/1600.
- Coverage line: "written-shape check (7d): N turns checked, M asked, K left
  unjudged", from the Stop hook's recorded outcomes; silent until the
  question has been put.
- test_guidance_ownership pins the new topic on reusing-code.
- prior-art hook header no longer says it stamps instance rows.

Plugin version minted.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

src/scribe/mcp/server.py

@@ -56,7 +56,8 @@ reads Agent Skills) and in each tool's description.
   matched, not none. Rules bind; preferences guide and you keep them current;
   lessons inform.
 - Search before acting or building, scoped with the active project_id; start
-  from a recorded snippet, and create_snippet what you build.
+  from a recorded snippet; before the turn ends, say what you built
+  (create_snippet the reusable, classify_shapes the rest).
 - Work is tasks (a fix is kind="issue"): in_progress on start, add_task_log as
   you go, done on finish; tag system_ids.
 - A plan is a milestone: find the existing one

src/scribe/services/coverage.py

@@ -922,6 +922,16 @@ async def compute_coverage(
     except Exception:
         logger.warning("stamp review read failed", exc_info=True)
         review = {}
+    # The write-time figure (milestone 439): how often the end-of-turn
+    # question was put on this project, and how often it was walked past.
+    # An audit reads the ledger; this reads whether the reflex held.
+    try:
+        from scribe.services import shape_check as shape_check_svc
+
+        write_time = await shape_check_svc.window_summary(project_id)
+    except Exception:
+        logger.warning("write-time figure read failed", exc_info=True)
+        write_time = None
     return {
         "total": len(rows),
         "accounted": len(rows) - unclassified,
@@ -950,6 +960,8 @@ async def compute_coverage(
         # holds (#4608).
         "weak_stamps": review.get("weak_count", 0),
         "incoherent_canons": review.get("incoherent_count", 0),
+        # {checked, asked, left, days} from the Stop hook's recorded outcomes.
+        "write_time": write_time,
         # Honesty flag, not decoration: every surface that shows the number
         # is expected to carry it through.
         "estimate": True,
@@ -1154,4 +1166,14 @@ def coverage_line(coverage: dict) -> str:
         review.append(f"{n_i} incoherent canon{'s' if n_i != 1 else ''}")
     if review:
         line += f"; {' · '.join(review)} to judge — stamps_to_review"
+    # Whether the writers judged what they wrote (milestone 439). Silent until
+    # the question has been put at least once: a project nobody has written to
+    # through the plugin has no figure, not a perfect one.
+    wt = coverage.get("write_time") or {}
+    if wt.get("checked"):
+        line += (
+            f"; written-shape check ({wt.get('days', 7)}d): {wt['checked']} "
+            f"turn{'s' if wt['checked'] != 1 else ''} checked, {wt.get('asked', 0)} asked, "
+            f"{wt.get('left', 0)} left unjudged"
+        )
     return line

src/scribe/services/shape_check.py

@@ -27,6 +27,9 @@ report_check gives.
 from __future__ import annotations
 
 import json
+from datetime import datetime, timedelta, timezone
+
+from sqlalchemy import select
 
 from scribe.models import async_session
 from scribe.models.app_log import AppLog
@@ -145,3 +148,45 @@ def parse_written(raw: str, *, cap: int = 200) -> list[tuple[str, str, str]]:
         if len(out) >= cap:
             break
     return out
+
+
+# The window the coverage line reports the write-time figure over. A week is
+# long enough to hold a few working sessions on a project and short enough
+# that a reflex that started slipping shows up while it is still news.
+WINDOW_DAYS = 7
+
+
+def summarise(outcomes: list[str]) -> dict:
+    """{checked, asked, left} from a list of recorded outcomes.
+
+    `checked` is turns the question was put to (a first stop that wrote
+    something): passed + blocked. `asked` is the blocked ones. `left` is
+    asks the agent walked past — the stop after a block that still had
+    unjudged shapes. That last number is the slip the milestone exists to
+    make visible."""
+    checked = sum(1 for o in outcomes if o in ("passed", "blocked"))
+    asked = sum(1 for o in outcomes if o == "blocked")
+    left = sum(1 for o in outcomes if o == "left_after_block")
+    return {"checked": checked, "asked": asked, "left": left}
+
+
+async def window_summary(project_id: int, *, days: int = WINDOW_DAYS) -> dict:
+    """The write-time figure for one project over the last ``days``."""
+    since = datetime.now(timezone.utc) - timedelta(days=days)
+    async with async_session() as session:
+        rows = (await session.execute(
+            select(AppLog.details).where(
+                AppLog.category == "plugin",
+                AppLog.action == "shape_check",
+                AppLog.created_at >= since,
+            )
+        )).scalars().all()
+    outcomes: list[str] = []
+    for raw in rows:
+        try:
+            details = json.loads(raw or "{}")
+        except (TypeError, ValueError):
+            continue
+        if details.get("project_id") == project_id:
+            outcomes.append(details.get("outcome") or "")
+    return {**summarise(outcomes), "days": days}