@clize/clize 0.35.0 → 0.35.2

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@clize/clize",
-  "version": "0.35.0",
+  "version": "0.35.2",
   "mcpName": "ai.clize/clize",
   "description": "The real-world capability layer for AI agents — domains, email, deploy, media, SEO/GEO. CLI + MCP.",
   "keywords": [

package/skills/clize-seo/SKILL.md

@@ -120,9 +120,10 @@ your list. No amount of enumeration produces this. Fold it back into the matrix
 
 ---
 
-## 2. The shape of one analysis
+## 2. The shape of one round
 
-A cold start is a funnel, and the funnel has a known size and a known price.
+A round is a funnel with a known size and a known price — and there is no such thing as an
+"analysis round": every round ends in pages or in a blocker the human must clear (§10).
 
 ```
 30–100 candidates            (§1: competitor teardown + capability matrix)
@@ -225,9 +226,24 @@ will — it cannot measure those. That half of the state exists only here.
   had recorded in prose that `invoice data extraction` was the site's third-best earning URL
   despite an enterprise-looking SERP; the next day's session opened a fresh ledger, re-derived
   the enterprise read from the SERP, and marked the page for retargeting. The rule that would
-  have caught it is in §4 ("Search Console beats the SERP read").
+  have caught it is in §4 ("Search Console beats the SERP read"). Import **skip rows and
+  cell-level lessons** only. Do not import site-level verdicts — "the content side has nothing
+  left", "the bottleneck is authority" — those expire, and are re-derived every round from the
+  cell count (§10, cadence); kunavo carried one such verdict through seven rounds, each session
+  re-confirming it instead of counting cells. And a family described in prose is not in the
+  ledger: a cluster, a locale axis or a page family with no rows and no `target page` will never
+  be built — write the rows (kunavo's nine-cell integrations family sat as a paragraph for two
+  rounds while every session executed "the existing ledger").
 - **Narrative goes underneath as footnotes.** Two or three lines about what the pattern means.
   Not six paragraphs of themes.
+- **Every ledger file counts, and a plan is a queue.** A site may carry more than one ledger
+  (tabledi: `worksheet.md` for en-US and `worksheet-global.md` for the other markets); rows in
+  all of them are the site's rows, and a round that counts only one file will report "nothing
+  to build" beside 27 verified cells. A batch plan written under an earlier count rule (`≤5,
+  wait for impressions`) is not re-planned and never "the next batch": each cell is re-gated
+  on its own SERP verdict, and every cell that passes is a `to build` row for step ⓪ of the
+  next session (§10). tabledi's B1–B4 — 27 locale tool cells, 15 SERPs bought — sat under
+  "next batch" through three sessions.
 
 **Page batches — the second table in the ledger.** Keyword rows hold bets; the batch table holds
 what the site's composition is doing, and it is what the batch gate and the retire rule (§7.3)
@@ -245,7 +261,7 @@ sitemap — and `unseenTotal` says how much of the site the list is showing you.
 
 ### The pass/fail test
 
-Six ways an analysis fails. The first is about *what you found*, the last two about *what you
+Six ways a round fails. The first is about *what you found*, the last two about *what you
 funded*; the middle three are about *whether you actually ran the analysis you claim to have
 run* — all six have been observed in real runs that read as finished.
 
@@ -273,8 +289,11 @@ run* — all six have been observed in real runs that read as finished.
    discussions visible in `items` — so this costs nothing to fix: write them as rows next to
    the pages they support. Observed in the field (2026-08-31 second run): five funded pages,
    placement nearly invisible in the deliverable.
-6. **Rows without pages.** Buildable rows — a serp verdict that is not a wall, an intent you can
-   serve — left unshipped at session end are stock, not output. The deliverable counts shipped
+6. **Rows without pages — and families without rows.** Buildable rows — a serp verdict that is
+   not a wall, an intent you can serve — left unshipped at session end, or handed to "next
+   round" without one of the three blockers in §10 ⑤, are stock, not output.
+   An opportunity written in prose (a cluster, a family, a locale axis) with no rows behind it
+   is test 1 wearing a different coat. The deliverable counts shipped
    URLs, not qualified rows; a ledger of twenty rows and two pages is a session that spent itself
    on the analysis. Observed on our own tenant zero: eleven weeks of rounds that produced ledgers
    and no pages. If a page costs a session, the first deliverable is the generator (§7).
@@ -527,6 +546,18 @@ each after the first. clize.ai went from one page per session to seven in one se
 a family generator existed, and tabledi shipped forty in two days from one. Build the generator
 before the second page of any family, and never machine-translate a family page.
 
+**On a site that already has tools, this round's page count is the locale siblings of the tools
+it has.** A sibling is one strings entry (tabledi: a `DICT` table in `tools/apps.js`; clize.ai:
+a `strings.<loc>.mjs`) and its gate is its own SERP (§7.2). Count existing tools × verified
+locale cells before anything else — tabledi had 23 tools, 6 markets, 27 cells already verified
+open, and three sessions wrote "next batch" beside them while building one page. New tools are
+code and tests, one to three per session: they are the next batch's rows, not this batch's
+excuse. And **building a page that passed the gates is never a question for the human.** The
+human's gates are money, external-facing changes (a redirect, DNS, a published post) and sends.
+Ask for those in one line and build while you wait — never bundle the pages into the question
+(tabledi, 2026-09-02: one redirect that needed a yes was asked together with three pages that
+did not, and all four waited).
+
 ### 7.1 The pipeline — five steps, each with a pass/fail
 
 | step | act | it passed if |
@@ -698,7 +729,9 @@ comparison with the ledger's `target page` (§3):
 ### `gsc.pages` — the page face
 
 The question the product exists for is "how much traffic does the final page get", answered by
-page since 0.34.0 — three facts side by side, no judgment:
+page since 0.34.0 — three facts side by side, no judgment. Reading it is the first fifteen
+minutes of a round, not the round: the shapes below say what to fix, and the round still
+enumerates and builds (§10).
 
 | list | what is in it | the facts per row |
 | --- | --- | --- |
@@ -773,36 +806,35 @@ sentence (`authority_limited`); the roundups still come from a deliberate `seo s
 
 ## 10. The cycle
 
-**The loop starts at measurement, and its purpose is allocation.** Two facts force the entry
-point. First, the only thing that *changes* between rounds is the measurement: sv/kd are
-monthly-refresh data (cached 30 days) and SERPs move slowly — re-run discovery two weeks later
-and it returns the same conclusions at the same prices. What changed is what *your* actions
-did to the world, and that shows up only in `check`, which is free. Second, the entry read is
-the branch point: each round's only real question is **where the next unit of work goes** —
-double down where a return signal appeared, stop where none did, discover only when the read
-hands you something new. Measurement is the allocator, not a dashboard.
-
-**A round is four steps: ① read the feedback (`check`, free) → ② allocate (double down / stop
-loss / new bet) → ③ execute (pages, placement, the occasional paid query) → ④ book it (write
-what you put in back into the ledger, §3).** Step ④ is the one that gets skipped, and skipping
-it breaks step ② two weeks later: "flat, no signal" and "flat, no signal, *and we already
-pitched five roundups*" are different rows leading to different decisions, and nothing but the
-ledger can tell them apart.
+**The loop starts at measurement, and its purpose is allocation.** Between rounds the only
+thing that changes is what *your* actions did to the world, and that shows up only in `check`,
+which is free (sv/kd are monthly data, SERPs move slowly). The entry read is the branch point:
+each round's real question is **where the next unit of work goes** — double down where a signal
+appeared, stop where none did, and keep enumerating cells the ledger has not priced yet.
+
+**A round is a pipeline, and it opens by building.** ⓪ **ship the qualified batch** — every
+`to build` row from the last round and every plan cell with its own SERP verdict (§3), in every
+ledger file — before reading a single number, while the session's budget is whole → ① read the
+feedback (`check`, free) → ② allocate (double down / stop loss / new bet) → ③ execute what is
+cheap now (siblings, retargets, placement drafts) and write the rest as the next session's
+`to build` rows → ④ book it (§3) → ⑤ report it (two tables, below). This round's analysis
+qualifies the *next* round's batch; it does not postpone this one's. Step ⓪ is the step that
+vanished in three consecutive runs: each read, diagnosed, wrote a ledger, named itself an
+"analysis round" and handed the pages to "next time". Step ④ is the other one that gets
+skipped, and skipping it breaks step ② two weeks later: "flat, no signal" and "flat, no signal,
+*and we already pitched five roundups*" are different rows, and only the ledger can tell them
+apart.
 
 ```
 clize seo check            # free — every session on an existing site starts here
 ```
 
 **Domain of the loop: a site that exists.** `check` needs `--domain` (or a clize project
-directory) — with neither it errors, correctly. No site yet, or a brand-new venture with
-nothing deployed? That is pre-cycle work: run discovery (§1) to decide what to build, ship
-the first batch (§7), enroll the keywords — the cycle starts at that first check.
-
-**Round one is not special — it is the round that reads zero.** The zeros are your baseline,
-and the entry read still pays before you spend a cent: it surfaces the GSC service-account
-setup (a 2-minute human step — hand it to your human now, so it runs in parallel with the
-analysis), and on a site with any history it hands you free seeds (`gsc.topQueries` /
-`newQueries`) before you buy a single word.
+directory). No site yet? That is pre-cycle work: discovery (§1), first batch (§7), enroll — the
+cycle starts at that first check. **Round one is not special — it is the round that reads
+zero.** The zeros are the baseline; the entry read still pays before you spend a cent: it
+surfaces the GSC service-account setup (a 2-minute human step — hand it over now, in parallel),
+and on a site with history it hands you free seeds (`gsc.topQueries` / `newQueries`).
 
 ### ②–③ Allocate, then execute — branch on what the read says
 
@@ -816,7 +848,12 @@ analysis), and on a site with any history it hands you free seeds (`gsc.topQueri
 | `rank.pending` non-empty | those words are not measured yet — no branch applies to them this round; do not read their zero as "missed" |
 | `authority_limited` | placement only **on that theme** — another page on it will not move it (§6). The label is per keyword, not per site: production continues on every other theme. If every funded theme reads this way, the site is Shape C (top of this file): re-derive the ledger toward tool, long-tail and locale rows before pitching |
 | `gsc.newQueries` non-empty | mini-discovery: fold them into the matrix, price the new cells, extend the tracked list if they survive the SERP check |
-| no movers two rounds running *and* the placement queue is empty | the plan is stale, not slow — back to §1 and re-derive; more waiting will not convert a flat line |
+| no movers two rounds running | the plan is stale, not slow — back to §1 and re-derive toward new page types; more waiting will not convert a flat line. The placement queue does not gate this: a queue that has waited on a human for two rounds is recorded as `blocked:` and handed over, and production goes on without it — placement is human-gated, building is not, and one must never hold the other (kunavo's four ready pitches held production for seven rounds) |
+
+**Rows are not exclusive.** A round takes every branch its read matches — a technical fix, a
+placement action and a production batch in the same session is the normal shape, not three
+rounds. `technical round` and `indexing round` say what to do *for those pages*; they never
+cancel the production step for the rest of the site.
 
 ### ④ Book the round
 
@@ -834,17 +871,45 @@ next round's allocation possible:
   `gsc.pages.unseen` (found by `firstSeen`), and whether this window was another zero. This is
   the row the next round's batch gate and retire rule read.
 
-The stop-loss branch above ("no movers two rounds running *and* the placement queue is empty")
-is answerable only if the previous round wrote down what it put in. The product does not track
-your side of the bet, deliberately: it cannot measure whether a pitch was sent, and a field it
-cannot measure is just a form for you to fill in. That state lives in the ledger or nowhere.
+The stop-loss branch above ("no movers two rounds running") is answerable only if the previous
+round wrote down what it put in. The product does not track your side of the bet, deliberately:
+it cannot measure whether a pitch was sent, and a field it cannot measure is just a form for you
+to fill in. That state lives in the ledger or nowhere.
+
+### ⑤ Report the round
+
+What the human reads ends with two tables, not with findings. First, **this round's pages** —
+every row the round touched, one line each:
+
+| keyword | locale | serp | target URL | type | status |
+| --- | --- | --- | --- | --- | --- |
+| hreflang 生成 | ja-JP | open (small sites) | /ja/seo/hreflang-tag-generator/ | tool | built |
+| open webui api key | en-US | open | /docs/integrations/open-webui | integration doc | to build: generator first |
+| claude code preise | de-DE | official_wall | — | — | skip |
+
+`status` is one of `built` (shipped this round), `to build`, `exists` (a page already carries
+it), `skip` (with the reason). A `to build` row carries a blocker, and only three exist:
+`needs human` (a redirect, a send, money — asked in one line, §7), `needs tool code` (a new
+client-side tool; name the session it ships in), `needs generator` (no family scaffold yet).
+"Analysis round", "next batch" and "awaiting go-ahead" are not blockers — a row wearing one of
+those means the session is not finished, and it goes back to step ⓪. Second, **placement** — target /
+page / status (`sent` / `drafted` / `blocked: needs human`). Then at most five lines of
+narrative. A report that is findings and no pages table is test 1 (§3) in its final form: the
+human cannot act on a paragraph. A round with an empty pages table says so *in* the table, with
+the cell count that justifies it — kunavo round five delivered four findings, four
+opportunities and no table, and the three buildable families in its own prose went unbuilt.
 
 **Cadence**: SEO is a weeks-to-months system and the window `check` measures is 4 weeks wide —
-checking daily just re-reads the same window. Every two weeks is plenty. And a round that
-re-reads an unchanged window must not re-decide: if `check` shows nothing new since the last
-round (same window, no movers, no new queries), the round is execution of the existing ledger,
-not a new analysis. tabledi 2026-09-02 re-ran discovery on the same data as 09-01 and flipped
-its P0 — a recommendation that flips on identical data is not a reading, it is noise.
+checking daily just re-reads the same window. Every two weeks is plenty. A round that re-reads
+an unchanged window must not re-judge *existing rows*: same window, no movers, no new queries →
+the P0s stand (tabledi 2026-09-02 re-ran discovery on the same data as 09-01 and flipped its
+P0 — a recommendation that flips on identical data is noise). **That rule never means no new
+cells.** Pricing a cell the ledger has never priced is not re-reading the window — the window
+has no opinion on it. Every round, whatever the window: count the ledger's unbuilt buildable
+rows against the site's cell space (§1.2: 100–150 for a full site); fewer than ten → this round
+enumerates (§1) and builds. kunavo, 2026-09-02, round five: same window, ledger exhausted,
+three markets never priced, a nine-cell family in prose — the round shipped zero pages and
+deferred the pricing for the third time, quoting this paragraph.
 
 ---