chore: correct the roborev exclusions and refresh the guidelines

roborev matches excluded_commit_patterns as substrings, and compass's cadence rule
requires `type(ws): subject (#N)`, so `chore:` never matched `chore(engine): ...`.
Adds `chore(`.

Does not add `docs(`. Those commits carry the journal entry and the board
narrative, and the plain-language rule exists to check exactly that prose -- it has
no other commit to fire on.

The guidelines were also stale: they were composed before base.md gained the
plain-language rule, and nothing re-composes them on its own. Refreshed, which is
what put that rule in this repository for the first time.

.Rbuildignore

@@ -3,6 +3,7 @@
 ^\.Rproj\.user$
 ^_pkgdown\.yml$
 ^docs$
+^pm$
 ^Meta$
 ^doc$
 ^pkgdown$

.gitignore

@@ -4,17 +4,11 @@
 .Ruserdata
 *.Rproj
 inst/doc
-# pkgdown output. Listed as children rather than `docs/` so compass's
-# docs/pm/ and docs/decisions/ can be re-included -- git cannot re-include
-# anything beneath an excluded directory.
-#
-# Note the anchoring change this forces: a bare `docs/` matches a directory of
-# that name at ANY depth, while `/docs/*` matches only at the repo root. The
-# fixture corpus's own docs/ therefore needs its own rule to stay excluded.
-/docs/*
-!/docs/pm/
-!/docs/decisions/
-inst/extdata/fixture_corpus/docs/
+# pkgdown output. Compass used to keep its files in docs/pm/ and
+# docs/decisions/, which forced this to be written as children with two
+# re-includes -- git cannot re-include anything beneath an excluded directory.
+# Compass lives in pm/ now, so the whole directory can be excluded again.
+docs/
 /doc/
 /Meta/
 .DS_Store

.roborev.toml

@@ -2,7 +2,7 @@
 # Reviews are queued to a background daemon -- they never block a commit.
 
 post_commit_review = 'commit'
-excluded_commit_patterns = ['WIP', 'chore:', 'docs:', 'Merge ']
+excluded_commit_patterns = ['WIP', 'chore:', 'chore(', 'docs:', 'Merge ']
 
 review_guidelines = '''
 # --- compass:begin (generated -- edit the sources, not this) ---
@@ -19,6 +19,10 @@ review_guidelines = '''
   deeper than four levels.
 - No magic numbers or hardcoded paths -- name them as constants or read them from config.
 - New behaviour needs a test. A bug fix needs a test that fails without the fix.
+- Prose a person reads -- an issue title or body, a journal entry, a decision record,
+  the narrative on the status board -- names the action or the thing, not the shape of
+  the machinery. Flag "gate", "seam", "surface area", "load-bearing", "first-class",
+  "primitive", "blast radius". A project's own defined vocabulary is not the target.
 - Use the native pipe `|>`, not magrittr `%>%`.
 - snake_case for objects and functions; UPPER_SNAKE for constants. Never use `.` as a
   word separator in a function name -- it collides with S3 dispatch.

pm/STATUS.md

@@ -14,14 +14,23 @@ The six open issues split cleanly. Two are API work carried out of the #9 review
 and deliberately deferred there rather than fixed in that branch. Three concern the
 corpus layer, and the largest of them, partition-level caching, was named the single
 highest-leverage change on the remote path before being deferred. One, the
-data-correction intake, is a decision rather than a task: it was parked during the
-0.3.0 design and it gates the API announcement, because without it the corpus cannot
-make the "traceable and correctable" claim that most distinguishes it from Census's
-own files.
+data-correction intake (#52), is a decision rather than a task: it was parked during
+the 0.3.0 design, and the API announcement waits on it, because without it the corpus
+cannot make the "traceable and correctable" claim that most distinguishes it from
+Census's own files.
 
 Nothing here is blocked on anything else, so the ordering is a judgement about value
 rather than a dependency graph.
 
+Compass's own files moved out of `docs/` this session. They were sitting inside
+pkgdown's output directory, and `pkgdown::clean_site()` deletes every top-level entry
+there except `CNAME` and `dev` — asked directly, it listed `docs/pm` and
+`docs/decisions` among the 28 it would remove, with the guard that would have stopped
+it satisfied by `docs/pkgdown.yml`. They are in `pm/` now. Nothing was lost: the
+journal had no entries and there were no decision records yet, which made this the
+cheapest moment to move. The `.gitignore` workaround that re-included two children of
+an excluded `docs/` is gone with it.
+
 ## Ready to work on next
 
 - **#34** cog_revenue() offers expenditure recipes as suggestions: scope the candidate query by category_type · `ws/api` — nothing is blocking it; something is currently wrong
@@ -47,25 +56,10 @@ rather than a dependency graph.
 <details>
 <summary>Dependency graph and detail</summary>
 
-```mermaid
-graph TD
-  I34["#34 cog_revenue() offers expenditure recipes as sug…"]
-  I36["#36 n_units_reporting is category-conditional and c…"]
-  I2["#2 Extend population data to be households as an a…"]
-  I33["#33 Decompose .build_suggestions() (106 lines) into…"]
-  I52["#52 Release 11/11: design the data-correction intak…"]
-  I64["#64 Partition-level caching: R/cache.R is still a s…"]
-  class I34 ready;
-  class I36 ready;
-  class I2 ready;
-  class I33 ready;
-  class I52 ready;
-  class I64 ready;
-  classDef ready fill:#dafbe1,stroke:#2da44e;
-```
+_Nothing blocks anything else, so there is no graph to draw._
 
 - Marker: `none` (no journal entry yet)
-- Commits since: 164
+- Commits since: 165
 - Open issues: 6
 
 </details>