docs: correct the writing guide's counts and examples after the second fact check

With the script as it now counts bold-led paragraphs, the paper before the October pass
had 38 paragraphs over 100 words. Commit 29ff584 cut 1,200 words of body. The lead-in
example is the sentence commit 16fdc32 removed. The Why rule lets a subsection open with
its mechanism when its parent gave the reason. The milestone lists and "each has an
instance" are gone, and the incidents in section 1 are plural. In the paper, the section 8
example points at Promotion for curation, two count lead-ins before bold-led lists are
cut, and the section 9 sentence on who measures progress is three sentences.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
diff --git a/WRITING.md b/WRITING.md
index 4bb984b..d09ddd9 100644
--- a/WRITING.md
+++ b/WRITING.md
@@ -4,7 +4,7 @@
 below were recovered from the commits that shortened the paper in October 2026 and revised
 it in early November. Measured with the script at the end of this file, the body went from
 13,000 words at the commit before the pass (57d67d1, 12 October) to 8,700 at the end of the
-pass (29ff584, 13 October), and from 27 paragraphs over 100 words to none (29bf659, 15
+pass (29ff584, 13 October), and from 38 paragraphs over 100 words to none (29bf659, 15
 October). Most rules give one example from those commits, before and after, so that the
 next pass can tell the rule from a taste. The last section is the procedure a pass runs.
 
@@ -19,8 +19,8 @@
 > *Agent fleets need shared memory, and shared memory spreads mistakes and leaked secrets as easily as knowledge. It has to be cheap to read, safe to share, and outlast every model and tool it runs on. This paper proposes how.*
 
 **The section that introduces a part or a role opens with why it exists, labelled
-`**Why.**`** A subsection whose reason its parent gave may open with an example or a
-diagram instead. A section whose first paragraph is a mechanism has lost its reason; where
+`**Why.**`** A subsection whose reason its parent gave may open with its mechanism, an
+example or a diagram. A section whose first paragraph is a mechanism has lost its reason; where
 none can be written, the section may not belong.
 
 > − Organisations rename, split, merge, and hand names to new owners. ...
@@ -44,8 +44,7 @@
 > + Content moves to a lower label only through `declassify`, and a guard named in policy approves each one. ... Appendix E gives the rules in detail: who may attest handling, tool servers that are also sources, grants, and who guards which crossing.
 
 **Parallel facts go in a table with a Why column.** Three or more items that share the
-same two or three attributes are a table. A rules table has a Why column; a milestone has
-two lists, "Delivers" and "Proven when".
+same two or three attributes are a table. A rules table has a Why column.
 
 > | Rule | What it says | Why |
 > | **Leases** | A claim expires unless renewed | A crashed agent must never hold work hostage |
@@ -53,7 +52,7 @@
 **Cut what only repeats something said elsewhere.** For each section, name one thing a
 reader learns there and nowhere else. If nothing, cut it and redirect its references.
 
-> Build Order added little the evaluation and survival sections do not already say; remove it and point its few references at §11 instead. (commit 29ff584, which removed 1,600 words)
+> Build Order added little the evaluation and survival sections do not already say; remove it and point its few references at §11 instead. (commit 29ff584, which cut 1,200 words of body)
 
 **A term is explained where it first appears, and the glossary holds the definition.** The
 first occurrence carries an appositive ("the harness, the program that runs the model") or
@@ -123,7 +122,7 @@
 ## Tells
 
 These are removed on the lines a change adds. They are the marks of text written by a model
-or of a discussion restated as prose, and each has an instance from the commits.
+or of a discussion restated as prose, and most have an instance from the commits.
 
 **Contrast for effect.** "X, not Y", "rather than", "instead of", "never ... but". Delete
 the negative half; if the sentence still says the rule, the contrast was for effect. Keep a
@@ -143,8 +142,11 @@
 > − The caller states facts, not a length, and the decision is logged for replay.
 > + The decision is logged for replay.
 
-**Lead-ins that only count what follows.** "Three findings shape this section." is cut when
-the list has bold leads that carry it.
+**Lead-ins that only count what follows.** A sentence whose only content is a count and a
+promise is cut when the list has bold leads that carry it.
+
+> − Three findings from security and knowledge-system research shape this section.
+> (the three bullets follow directly from the **Why.**)
 
 **The aphorism and the closing flourish.** A short sentence at the end of a paragraph that
 restates it; a semicolon joining two clauses that mirror each other.
@@ -161,7 +163,7 @@
 **A company or product as the authority.** Named deployments appear in §13 and the
 references; the body cites them by kind ("one personal-workspace deployment") and every
 rule keeps its own reason. The body names a party only where the name is the fact: the
-incident in §1, a component the design adopts (OPA, OpenTelemetry), and Gas Town, the one
+incidents in §1, a component the design adopts (OPA, OpenTelemetry), and Gas Town, the one
 operator's fleet the paper cites most, which §4 defines once.
 
 > − At Meta a personal workspace of this kind ... reached 63,000 installs
diff --git a/docs/durable-memory-for-disposable-agents.md b/docs/durable-memory-for-disposable-agents.md
index d509f39..c43622a 100644
--- a/docs/durable-memory-for-disposable-agents.md
+++ b/docs/durable-memory-for-disposable-agents.md
@@ -553,8 +553,6 @@
 
 **An example.** Gas Town ([§4](#model-output-in-particular)) ran 450 rules that grew in three weeks to 650 refusal points, until no work was legal and it stopped. The reset that followed removed the gates people had set along with the ones agents had added [[117]](#ref-117), [[121]](#ref-121).
 
-Four rules follow.
-
 - **Enforced rules per domain have a ceiling** ([§12](#cost-and-throughput)). Each rule over it is enforced only after a person signs it off, one rule at a time; a signature admits that rule and does not raise the ceiling.
 - **A reset keeps every rule with a person in its lineage.** A reset is the domain's owner retiring its rules in one step, when the stack has grown past use. It keeps every rule written or promoted by a person and retires the rest.
 - **An unused rule goes back to its owner.** A rule that produced no denial and no citation in a review period goes to its owner for retirement.
@@ -734,7 +732,7 @@
 
 Builders of long-running agents carry state between sessions the same way, in progress notes and version history [[109]](#ref-109). A fresh start from one clear restatement can also do better than a long history, because models rely too heavily on their earlier, wrong attempts [[110]](#ref-110). Hospitals that adopted structured handoffs made fewer errors [[111]](#ref-111). But an LLM summary of a run can hide signs that an attempt is failing [[100]](#ref-100), so the next agent checks a handoff against the ledger and the work itself.
 
-Where workers' outputs meet, the orchestrator checks for overlapping artifacts, runs the policy gate, and queues the decisions made along the way for curation ([§5](#5-the-curator)). Every finished task is kept in the record with its edges, handoff, and outcome. That history is the ground truth for evaluation ([§11](#11-evaluating-change)).
+Where workers' outputs meet, the orchestrator checks for overlapping artifacts, runs the policy gate, and queues the decisions made along the way for curation ([§4](#promotion)). Every finished task is kept in the record with its edges, handoff, and outcome. That history is the ground truth for evaluation ([§11](#11-evaluating-change)).
 
 ## Taps
 
@@ -808,8 +806,6 @@
 
 A session that boots into a seat takes the seat's T0 slice, presents the manifest its agent definition fixes ([§10](#10-evolving-the-system)) narrowed by the seat's authority envelope, and takes the tasks the seat's charter covers. The seat grants nothing by itself: the perimeter checks every action against the seat's authority envelope, the task definition's declared tools, and the principal chain, and the narrowest applies.
 
-Three rules follow.
-
 - **Secondment.** A session under an agent definition other than the one the charter names, a substitute model for instance, may read the seat's memory and write to the task chain. It may not write the seat's history, commitments, or handoff cache, because those are what the next occupant will trust.
 - **Retirement.** A seat's owner retires it through the review gate. Its open commitments return to the ledger for matching, and its history stays in the record under its identifier.
 - **Waking.** When a task pinned to a seat with no live session is admitted, the service starts a session in the seat, under the agent definition its charter names, so a pin names a role and never a process.
@@ -883,7 +879,7 @@
 - known-item lookups by ID, title, or error string in the top three;
 - `claim_next` within one second at the 99th percentile.
 
-Two parties measure progress: the **evaluator**, the role that holds the withheld scenarios and grades completions, tuned apart from the sessions it grades ([§10](#10-evolving-the-system)), and the landing queue, which the fleet cannot edit and whose count is read after a fixed delay. A task an agent filed and completed itself counts as upkeep. A check an agent changed counts only once a person has reviewed the change. A fleet that files and grades its own work can show throughput without delivering anything [[126]](#ref-126).
+Two parties measure progress. The **evaluator** is the role that holds the withheld scenarios and grades completions, tuned apart from the sessions it grades ([§10](#10-evolving-the-system)). The landing queue is the other: the fleet cannot edit it, and its count is read after a fixed delay. A task an agent filed and completed itself counts as upkeep. A check an agent changed counts only once a person has reviewed the change. A fleet that files and grades its own work can show throughput without delivering anything [[126]](#ref-126).
 
 The spend behind the numbers is a product of five terms: