@haus-tech/haus-workflow 5.12.1 → 6.1.0

package/library/global/skills/request/SKILL.md

@@ -13,19 +13,19 @@ the respective skill. Only routing, context carrying, the archive and the dispat
 
 `/request [task]`: without an argument, show the menu below.
 
-| Task                                | Does                                                                                                                                                   |
-| ----------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ |
-| `speca <request or gid>`            | Source manifest → dispatch `request-specifier` → archive spec + reviewer block                                                                         |
-| `designa <gid, slug or nr>`         | Gate on `✅ Spec ok` → `superpowers:brainstorming` with the user → the design doc in the repo + `design.md` in the archive; never Asana                |
-| `bedöm <gid, slug or nr>`           | Gate → secure an analysis → dispatch `request-scorer` → archive JSON                                                                                   |
-| `status`                            | Queue state via request-sync, BLUF table with row numbers                                                                                              |
-| `godkänn <nr or gid>`               | Move Offer status forward via request-sync; see the gate section                                                                                       |
-| `underkänn <nr or gid> <reason>`    | Reason as an Asana comment + iteration counter via request-sync                                                                                        |
-| `döm <item>`                        | Human ruling on OLIKA: choose a run, order a rerun or reject the source material                                                                       |
-| `detaljer <gid, slug or nr> [step]` | Print archived JSON/documents for the item                                                                                                             |
-| `byt titel <nr or gid>`             | Rename the Asana task to the spec title (`state.json.title`) via request-sync, on request only                                                         |
-| `utvärdera <item…> <model…>`        | Run the flow once per named model against the experiment area, at least one model; `sammanställ` gives the sweep view, `rapport` a readable comparison |
-| `hjälp`                             | Explain the tasks, touch no files                                                                                                                      |
+| Task                                        | Does                                                                                                                                                                           |
+| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| `speca <request or gid>`                    | Source manifest → dispatch `request-specifier` → archive spec + reviewer block                                                                                                 |
+| `designa <gid, slug or nr>`                 | Gate on `✅ Spec ok` → `superpowers:brainstorming` with the user → the design doc in the repo + `design.md` in the archive; never Asana; a clean record continues into `bedöm` |
+| `bedöm <gid, slug or nr>`                   | Gate → form check → `bedom.mjs` prices (a design draft and its review in draft mode, analysis, two scorers, archive) → BLUF; with several items or none, a batch               |
+| `status`                                    | Queue state via request-sync, BLUF table with row numbers                                                                                                                      |
+| `godkänn <nr or gid>`                       | Move Offer status forward via request-sync; see the gate section                                                                                                               |
+| `underkänn <nr or gid> <reason>`            | Reason as an Asana comment + iteration counter via request-sync                                                                                                                |
+| `döm <item>`                                | Human ruling on OLIKA: choose a run, order a rerun or reject the source material                                                                                               |
+| `detaljer <gid, slug or nr> [del-N] [step]` | Print archived JSON/documents for the item, or for one split part                                                                                                              |
+| `byt titel <nr or gid>`                     | Rename the Asana task to the spec title (`state.json.title`) via request-sync, on request only                                                                                 |
+| `utvärdera <item…> <model…>`                | Run the flow once per named model against the experiment area, at least one model; `sammanställ` gives the sweep view, `rapport` a readable comparison                         |
+| `hjälp`                                     | Explain the tasks, touch no files                                                                                                                                              |
 
 ## Hard rules
 
@@ -123,8 +123,18 @@ this skill (and the agents through their prompts) maintains it.
 the same request) becomes several directories: the score is set per item, so a directory
 with several specs cannot be assessed.
 
+The archive root is the main checkout's `.haus-workflow/requests/`, never a worktree's
+(ADR-0154). At the start of every task that reads or writes the archive, run
+`node <skillkatalog>/scripts/archive.mjs rot` first and use its `ARKIVROT:` line as
+`<arkivrot>` below; the cwd may be the main checkout or any of its worktrees.
+`IGNORERAD: nej` stops the task before any write: name `haus apply --write` in the main
+checkout as the fix (it writes the rule) and rerun `rot` once the user says it has run;
+never run it yourself from this skill. `rot` refuses a bare repository, a submodule
+checkout and a `--separate-git-dir` clone (their common dir is not named `.git`): run the
+task from the superproject's own checkout, or from an ordinary clone.
+
 ```
-<repo-rot>/.haus-workflow/requests/<gid-eller-slug>/
+<arkivrot>/<gid-eller-slug>/
 ├── spec.md                        ONE spec, byte-faithful: what the reviewer approves
 ├── granskarblock.md               "## For the reviewer" (bundle: the part's own
 │                                  block, never a copied one, ADR-0118)
@@ -156,7 +166,9 @@ its stop (`designa` step 5), with no runs row, since the design is no dispatch.
   the line yourself, and never hand-edit the block. If the archive is still tracked in git:
   run `haus apply --write`, which both writes the rule and untracks the files.
   Pricing material must not enter the customer's history. `.haus-workflow/request.json` (the sync's
-  config) stays tracked: it is configuration, not working material.
+  config) stays tracked: it is configuration, not working material. It lives in the main
+  checkout and carries the board's choices (`auto_sync_spec`, ADR-0157; `design_mode`,
+  ADR-0156).
 - **Old archive under `.haus/`:** moved by `haus apply` to
   `.haus-workflow/` (ADR-0103). Never write to `.haus/` again, and do not move it
   yourself: run `haus apply --write` and let the command do it.
@@ -179,6 +191,7 @@ its stop (`designa` step 5), with no runs row, since the design is no dispatch.
   "respec_varv": 0,
   "coverage_post": null,
   "design": null,
+  "board": null,
   "runs": []
 }
 ```
@@ -195,15 +208,19 @@ its stop (`designa` step 5), with no runs row, since the design is no dispatch.
   strings); `nasta-steg` and `bluf` refuse a split whose parts disagree on `of` or carry a
   non-integer `of`.
 - `stopp` = a stop outcome that the scripts have proven without a scorer:
-  `{ "utfall": "KONCEPT" | "DIAGNOS" | "TILLBAKA_TILL_SPEC" | "ESKALERA", "kalla": "spec.md" | "impact.md" | "granskarblock.md" | "design-avvisad-<datum>.md", "date": "YYYY-MM-DD", "orsak"?: string }`,
-  the last one for a design refused on `## Spec conflicts` (`designa` step 5).
+  `{ "utfall": "KONCEPT" | "DIAGNOS" | "TILLBAKA_TILL_SPEC" | "ESKALERA", "kalla": "spec.md" | "impact.md" | "granskarblock.md" | "design-avvisad-<datum>.md" | "design-intervju-kravs-<datum>.md", "date": "YYYY-MM-DD", "orsak"?: string }`,
+  `design-avvisad-<datum>.md` for a design refused on `## Spec conflicts` (`designa` step 5),
+  `design-intervju-kravs-<datum>.md` for the designer's `INTERVJU KRÄVS` in draft mode
+  (always ESKALERA; `add-design` clears it, since `designa` answers it).
   Written ONLY by `archive.mjs stopp` (which itself checks that the source carries
   what is claimed: the candidate marker in `spec.md`, the label `contradiction` in
   `impact.md`, the price-decisive open choice in `granskarblock.md`, ADR-0118, at
-  least one conflict under `## Spec conflicts` in `design-avvisad-<datum>.md`).
+  least one conflict under `## Spec conflicts` in `design-avvisad-<datum>.md`, a first line
+  `INTERVJU KRÄVS: <reason>` in `design-intervju-kravs-<datum>.md`).
   `add-design` clears a design-conflict stop (`kalla` `impact.md`, `orsak`
   `design conflict: …`) and supersedes the analysis it came from (`impact.md` →
-  `impact-superseded-<datum>.md`), or a receipt replaces it (`designa` step 5); every
+  `impact-superseded-<datum>.md`), or a receipt replaces it (`designa` step 5); it also
+  clears a stop from `design-intervju-kravs-<datum>.md`, which only a design answers; every
   other stop is reset ONLY by `archive.mjs respec`.
   `orsak` is set when `utfall` is `ESKALERA` (the command requires `--orsak` then,
   so a human sees why the item did not go back to the spec once more) and for the
@@ -219,10 +236,14 @@ its stop (`designa` step 5), with no runs row, since the design is no dispatch.
   later): an item never goes a third spec round on the same choice. Older archives
   lack the field and are read as `0`.
 - `coverage_post` mirrors the score envelope's optional `coverage_post: { yta, omfattning }`
-  (ADR-0119). `archive.mjs add-run` copies the field from the
-  latest pricing file that carries it, the skeleton (`ny-korning`/`ny-del`) sets
-  it to `null`. `bluf` shows `täckningspost: <yta>` in the `anm` column when
-  the field is set.
+  (ADR-0119) from the score pricing file that counts, the same file `bluf` reads the score
+  from: the round's main file, or the file a ruling's `galler_fil` names (ADR-0159).
+  `archive.mjs` recomputes it on `add-run` of a score pricing file, on `utslag` on the
+  score axis and on `set-status priced`, and sets it to `null` when that file has no
+  post or is not written yet; `respec` clears it. A `-run2` file sets it exactly when a
+  ruling makes it the counting file. Never write the field by hand. The skeleton
+  (`ny-korning`/`ny-del`) sets it to `null`. `bluf` shows `täckningspost: <yta>` in the
+  `anm` column when the field is set.
 - `design` is written ONLY by `archive.mjs add-design` and reset ONLY by
   `archive.mjs respec`: `{ "path": "<absolute path to the canonical file in the repo>",
  "file": "design.md", "sha256": "<marker-stripped hash of design.md, identical to the canonical file>",
@@ -231,7 +252,23 @@ its stop (`designa` step 5), with no runs row, since the design is no dispatch.
   `add-design` writes no runs row: `runs[]` records dispatches and the design is not
   one; the receipt is the field itself, write-once per version (a new design puts the
   old copy aside as `design-superseded-<datum>.md`). Older archives lack the field
-  and are read as `null`.
+  and are read as `null`. The field may also carry
+  `review` = `{ verdict, sha256, date, model }`, written ONLY by `archive.mjs add-design-review`
+  (the second reader's verdict on a draft), reset by `add-design` and `respec`
+  (both put `design-review.md` aside as `design-review-superseded-<datum>.md`);
+  `precheck --design` refuses to price a `stopp`. Whether the design is a draft is read
+  from `mode: draft` in `design.md`'s frontmatter, never from `state.json` (ADR-0156).
+- `board` is the receipt of the last Offer status move this flow recorded: speca step 7, bedöm step 8 (on a split root, the move of the parent) and `underkänn`; `godkänn` and `byt titel` record none. Form `{ "offer_status": "<value>", "date": "YYYY-MM-DD" }`, written ONLY by `archive.mjs synced` and reset to `null` by `archive.mjs respec` and by `add-run` on a round's main pricing file. On a split root it is the receipt of the sync that moved the parent; `respec` or a new round's main pricing file on one of its parts resets the root's receipt too; a cache note, never the truth (Hard rules).
+- `korning` = the running receipt `{ "pid": <heltal>, "started": "<ISO>", "katalog": "<abs>|null" }`
+  or `null`: written ONLY by `archive.mjs korning-start` (under the archive lock; it
+  refuses a living pid, so two starters never price the same item) and cleared ONLY by
+  `archive.mjs korning-slut`. `nasta-steg` writes `pågår (pid N)` while the pid lives and
+  `avbruten körning` when it is dead; the next `korning-start` takes a dead pid over;
+  `respec` refuses while a pid lives. Older archives lack the field and are read as
+  `null` (ADR-0155).
+- a runs row may carry `repo_sha`, the `HEAD` of the checkout the assessor read, written
+  by `korning.mjs` through `add-run --repo-sha` on `impact.md` rows; absent on older rows
+  and on rows the main conversation stamps (ADR-0155).
 - `umbrella` is written ONLY by `archive.mjs paraply` (ADR-0144):
   `{ "date": "YYYY-MM-DD", "sha256": "<marker-stripped hash of the umbrella spec.md>" }`,
   the receipt that the parent's `spec.md` was rewritten one level up after a split.
@@ -239,7 +276,7 @@ its stop (`designa` step 5), with no runs row, since the design is no dispatch.
   parts keep their inherited copies of it. `paraply` also sets `title` from the umbrella's
   heading; approval, design, runs, stops and the parts are untouched. The field is absent
   until `paraply` writes it and is never reset.
-- `approval` is written by the gate in `designa` or `bedöm`:
+- `approval` is written ONLY by `archive.mjs approve`, run by the gate in `designa` or `bedöm` after it has read the board, and copied unchanged into a new split part by `archive.mjs ny-del` (the part inherits the approved spec, ADR-0144):
   `{ "source": "asana" | "manual", "by": "<namn, krävs vid manual>",
  "status_read": "<värde>", "date": "YYYY-MM-DD" }`.
   The note is traceability, never the approval itself.
@@ -321,7 +358,7 @@ its stop (`designa` step 5), with no runs row, since the design is no dispatch.
 Model runs from `utvärdera` live in a tree of their own and are never pricing items:
 
 ```
-<repo-rot>/.haus-workflow/requests/experiments/<slug>/
+<arkivrot>/experiments/<slug>/
 ├── experiment.json          ticket reference, date, model list, manifest hash,
 │                            status per run
 ├── <modell>/                `default` as the model value gives the directory name
@@ -403,6 +440,8 @@ Model runs from `utvärdera` live in a tree of their own and are never pricing i
    the heading skeleton in `RUBRIKER:` follows that language, see the prompt template) and
    `DEPS:` (`installerade` when `node_modules/` or equivalent exists under
    REPO-ROT, otherwise `saknas`; without a repo always `saknas`).
+   `REPO-ROT:` is the checkout whose code is read; it is not the archive root, which
+   `rot` gives.
    Write the manifest to `<katalog>/kallmanifest.md` in every directory the spec
    is archived in (step 5), so that `bedöm` can give the assessor the same manifest without
    building a new one. The manifest's first line is `SPRÅK: <svenska|engelska>`,
@@ -490,7 +529,13 @@ på request-write-spec/references/specformat.md>`. The script sets
    stop nothing.
 7. **Reply BLUF:** one line per spec (heading + any flags), then the reviewer block
    in its entirety. Offer to place the specs on the board via request-sync (🔍); do
-   it only on explicit request. Placing a spec on the board is no rename: request-sync
+   it only on explicit request when the main checkout's `.haus-workflow/request.json`
+   (`ARKIVROT:`'s parent) carries `auto_sync_spec: false`; otherwise (`true` or absent,
+   ADR-0157) place every spec of this round that passed the form check in step 6 on the board now, in this turn, in the
+   main conversation, one item at a time through request-sync's ordinary spec sync (the
+   Decision read, the read-back, the three automatic rename cases are request-sync's and do
+   not change), a candidate included (the customer dialogue is its next step), a spec with a BRIST or a standing stop (`PRISAVGORANDE`, ADR-0118) never, it waits for `synka` as today. The ⚖️ sync is not covered: it waits for `synka` (`bedöm` step 8, ADR-0155).
+   Placing a spec on the board is no rename: request-sync
    renames the task to the spec title by itself only in the three automatic cases
    (ADR-0144): the task name is empty, `Original request title` is empty, or the task is
    one of several minted from one source. In the Closing, under the board line,
@@ -498,30 +543,49 @@ på request-write-spec/references/specformat.md>`. The script sets
    title and `byt titel <item>` is the optional next action. When an item is placed on the board: set status
    `spec_synced`, rename the slug directory to the minted gid, note the slug
    in `former_slug` and update `request_id` in `state.json`.
+   Then record the write: `node <skillkatalog>/scripts/archive.mjs synced <katalog> --offer-status "<the status request-sync set>"`.
    **Close per `## Closing: where the item stands`** (the table from
-   `archive.mjs nasta-steg`, the board line, the ISÄR check), never without.
+   `archive.mjs nasta-steg`, the board line, the ISÄR check, and the wrap-up lines at a
+   resting point), never without.
 
 ## Task: `designa`
 
 Runs the design leg: brainstorming with the user against the approved spec, the
 design doc saved in the customer repo and recorded in the archive. Runs in the main
 conversation (the brainstorming skill asks the user questions), in a session whose cwd
-is the customer repo. Nothing here touches Asana: the design is internal source
+is the customer repo or one of its worktrees (the archive root comes from `rot`). Nothing here touches Asana: the design is internal source
 material for `bedöm`'s assessor and never reaches the board.
 
+**Draft mode (ADR-0156).** When `.haus-workflow/request.json` in the main checkout carries
+`design_mode: draft`, an approved item without a design is designed by the machine inside
+`bedöm`: `bedom.mjs` passes the mode on, `korning.mjs` dispatches `request-designer` (opus),
+records the doc with `add-design`, has `request-design-reviewer` (fable) read it, and goes on
+to the analysis only on `VERDICT: fortsätt`. `designa <item>` is still the interactive path
+in both modes and supersedes a draft. A draft carries `mode: draft` in its frontmatter and
+three sections the interactive doc may omit: `## Assumptions` (every HOW question the machine
+answered, the option taken, the options not taken), `## Removed or renamed` (every symbol
+the design deletes or renames; `precheck --design` and `add-design` grep each one outside
+the files under `## Files touched`, a hit refuses the record) and `## Files touched`. An
+architectural item comes back as `intervju krävs` with the receipt
+`design-intervju-kravs-<datum>.md` and an ESKALERA stop that `designa` clears, and a reviewer
+`stopp` leaves the item `designed` with `design-review.md` in the archive; both route here,
+to the interview.
+
 1. **Gate. Identical to `bedöm` step 1.** Read Offer status via request-sync in the
    main conversation immediately before the run. Only `✅ Spec ok` (normalized on the
    alphabetic core) lets it through. An item without a gid follows `bedöm`'s rules for
    that case (refuse when the repo has `.haus-workflow/request.json`, otherwise the
    explicit "approved by whom" question). An item with a gid but no local archive
-   directory: not supported, say so and stop. Note the passed gate in
-   `state.json.approval` exactly as `bedöm` does; a later `bedöm` rereads the board.
+   directory: not supported, say so and stop. Record the passed gate with
+   `archive.mjs approve` exactly as `bedöm` does; a later `bedöm` rereads the board.
 2. **Form check.** Run `node <skillkatalog>/scripts/precheck.mjs <katalog>`. A
    `KANDIDAT:` line or a standing `stopp` means the item is not ready for a design,
    with one exception: a stop from `impact.md` whose `orsak` starts with
    `design conflict:` (the assessor found the design unbuildable) is exactly what a
    redesign answers, continue to step 3; `add-design` clears that stop and puts the
    old analysis aside as `impact-superseded-<datum>.md` when it records the redesign.
+   The same holds for an ESKALERA stop from `design-intervju-kravs-<datum>.md` (the
+   designer asked for this interview): continue, `add-design` clears it.
    Any other such line or stop: route it as the stop outcome (see `bedöm`) and stop. A
    `spec.md` with more than one spec block is a BRIST, as in `bedöm`. Other BRISTER:
    fix the link that is defective before anything is designed.
@@ -530,8 +594,9 @@ material for `bedöm`'s assessor and never reaches the board.
    Otherwise, if `state.json.design` is set, run
    `node <skillkatalog>/scripts/precheck.mjs <katalog> --design`. Three branches:
    (a) it passes (the hash still matches `design.md` and the canonical file): ask ONE
-   question via AskUserQuestion: keep it, or redesign. Keep: close per step 6 without
-   a new brainstorming. Redesign: continue; `add-design` supersedes the old one
+   question via AskUserQuestion: keep it, or redesign. Keep: continue per step 6 without
+   a new brainstorming; a kept design that passed `precheck --design` continues into
+   `bedöm` the same way a fresh record does. Redesign: continue; `add-design` supersedes the old one
    (`design-superseded-<datum>.md`), it never overwrites.
    (b) its only design BRISTER are the hash lines
    (`design.md: hashen är inte state.design.sha256 …` and/or
@@ -604,12 +669,7 @@ material for `bedöm`'s assessor and never reaches the board.
    item's state (status `priced` or `guard_run`, a standing stop other than a design
    conflict) is not fixed in the doc: `respec`, or decide the stop first.
 
-6. **Reply BLUF:** one line per item (the design's title and path), the
-   `## Spec conflicts` section verbatim when it is not `none`, the commit offer, then
-   **Close per `## Closing: where the item stands`** (the table from
-   `archive.mjs nasta-steg`, the board line, the ISÄR check), never without. For a doc
-   refused on conflicts, the Closing's part 4 presents the conflicts as the choice
-   (`## Avgjorda val`, `respec`, `speca <adress>`).
+6. **Continue into `bedöm`, or close.** When `add-design` recorded the doc, `precheck --design` wrote no BRIST and `## Spec conflicts` is `none`, run `bedöm <adress>` now, in the same turn, from its step 1 (the gate is reread from the board; the design is internal, so nothing reaches the customer before `bedöm`'s own sync rule, ADR-0153): the turn then ends with `bedöm`'s BLUF and Closing, never with a second block. If `bedöm` stops before its step 8 (its gate, a candidate, a contradiction), its Closing still carries the design doc's path and the commit offer. VARNING lines stop nothing: the chain runs. In every other case (a refused doc, a BRIST, an interview that ended without a file), **Reply BLUF:** one line per item (the design's title and path), the `## Spec conflicts` section verbatim when it is not `none`, the commit offer, then **Close per `## Closing: where the item stands`** (the table from `archive.mjs nasta-steg`, the board line, the ISÄR check), never without. For a doc refused on conflicts, the Closing's part 4 presents the conflicts as the choice (`## Avgjorda val`, `respec`, `speca <adress>`).
 
 ## Task: `bedöm`
 
@@ -634,7 +694,10 @@ drift guard.
    - **The item has a gid but lacks a local archive directory:** not supported yet
      (hydration from the board comes with the batch step); say so and stop;
      the reviewer block cannot be recreated from Asana.
-   - Note the passed gate in `state.json.approval` (the manual case with `by`).
+   - Record the passed gate with
+     `node <skillkatalog>/scripts/archive.mjs approve <katalog> --source asana --status "<Offer status as read>"`
+     (the manual case: `--source manual --by <name>`); the script refuses any status
+     other than ✅ Spec ok, so a gate that was not read cannot be noted.
      The next `bedöm` reads the status again.
    - **Gate 1b, the design (only when the assessor is to be dispatched).** When
      `impact.md` is missing, or precheck rejects it on source-material provenance
@@ -647,6 +710,15 @@ drift guard.
      with one exception: a split part (`state.split.parent` set) is not gated, the
      split copied the parent's `design.md` into it, and the assessor gets it as the
      `DESIGN` block.
+     In draft mode (`design_mode: draft` in `.haus-workflow/request.json` in the main
+     checkout, ADR-0156), an item whose only design lines are the no-design signature,
+     `DESIGN: saknas` and the two BRISTER `design.md: saknas` and
+     `state.json: fältet "design" är null`, is not refused: it goes on to step 2, and
+     `bedom.mjs` has the design drafted and reviewed before the analysis (`designa`, Draft
+     mode). A recorded design is gated as above in both modes, a draft included: an
+     interactive doc without the draft sections passes with VARNING lines only, and a
+     stopped review (`designgranskningen stoppade utkastet`) is a BRIST whose way on is
+     `designa <adress>`.
 2. **Form check.** Run `node <skillkatalog>/scripts/precheck.mjs <katalog> --impact`
    (without `--impact` if the analysis does not exist yet). A `spec.md` with more than one
    spec block is a BRIST: split the item per the archive rule before anything is dispatched.
@@ -660,6 +732,37 @@ drift guard.
    an identical repeat call is harmless), and route the item as the stop outcome
    below. No analysis and no score may ever be dispatched on a
    candidate.
+
+**Steps 3 to 7 run as ONE process in production (ADR-0155).** After step 2 has passed,
+run `node <skillkatalog>/scripts/bedom.mjs --repo-rot <abs> --arkiv <arkivrot> --post <gid>`
+as a background process and wait for its exit: it starts `korning.mjs` in production mode
+on the item directory, which runs the analysis when it is missing or stale, the two
+scorers at once, `kontroll.mjs`, `compare-scores.mjs` and the archiving through
+`archive.mjs`, exactly as steps 3 to 7 describe; those steps are the contract the script
+implements and its tests pin, not instructions for the main conversation, which
+dispatches no agent and runs no check in this leg. `--repo-rot` is the checkout whose
+code the assessor reads (the session's cwd; a worktree is fine) and `--arkiv` is
+`ARKIVROT:` from `rot`. The starter writes the running receipt (`archive.mjs korning-start`,
+`state.korning`) before the process starts and clears it after
+(`korning-slut`), skips an item whose receipt carries a living pid (`pågår`), and stops
+starting items when the batch's cost reaches `max_usd_per_batch` in
+`.haus-workflow/request.json` (default 100 USD; `--max-usd` overrides one run), counting an
+item in flight at the average of the finished ones. Read its
+final table (`| post | status | poäng | utfall | wall | USD |`), paste it under the BLUF
+in step 8, and route every row by its status: `priced` goes on to step 8, `stopp` and
+`väntar på dom` take the stop and OLIKA routing below, `ofärdig` shows the reason from
+the table and `<post>/korningar/<körning>/stderr.log`, `grind` and `tak nått` show the
+reason and stop, and a `pågår` row is left alone.
+`bedom.mjs` passes the board's design mode on to `korning.mjs` as `--design-mode`
+(`design_mode` in `.haus-workflow/request.json`; `interview` and an absent key read as
+`intervju`; the flag overrides one run). In draft mode the process drafts a missing design
+(`request-designer`, recorded with `add-design`) and has `request-design-reviewer` read it
+before the analysis (ADR-0156); the verdict is a judgment and is never redispatched. A
+`stopp` row whose reason starts with `intervju krävs:` (the stop from
+`design-intervju-kravs-<datum>.md`) or reads `designgranskningen stoppade (design-review.md)`
+routes to `designa <adress>`, the interview. Commit nothing in `--repo-rot` while
+the process runs: the impact row records the `HEAD` the assessor read (`repo_sha`).
+
 3. **Analysis.** If `impact.md` is missing, or precheck rejects it on
    source-material provenance (the analysis was made against another spec or another design): this requires a repo in
    the session (otherwise stop: a guessed analysis is forbidden by the skill itself).
@@ -692,9 +795,10 @@ drift guard.
    (the command sets status `assessed`), skip steps 4 to 7 and route
    the item as the stop outcome below: no scorer is dispatched on source material
    that already says the spec must go back. When the deviation line's text after the
-   label starts with `design conflict:`, pass `--orsak "<that text>"`: the item goes
-   back to the design, not to the spec, and the Closing routes it to `designa <adress>`
-   (a redesign clears the stop, see `add-design`). `ingen` and `precisering` go
+   label starts with `design conflict:`, the command reads that text from `impact.md` as
+   the stop's `orsak` (#595; an `--orsak` that contradicts the file is refused, exit 2):
+   the item goes back to the design, not to the spec, and the Closing routes it to
+   `designa <adress>` (a redesign clears the stop, see `add-design`). `ingen` and `precisering` go
    on to step 4. If precheck rejects the new
    draft → redispatch the assessor ONCE per the redispatch rule, and give
    the rejected draft a receipt exactly as in speca step 6: archive it as
@@ -718,8 +822,12 @@ drift guard.
    `resolve-anchors.mjs` (evidence anchors to verbatim source blocks;
    OLÖST/FLERTYDIG is a contract breach) and contract-check in one process; a
    protocol-term breach in a pure customer-text field is rewritten by a cheap model
-   instead of costing a new scorer run. The `REPARERAT:` lines are shown in
-   the BLUF reply and written into the runs row's `--note`.
+   instead of costing a new scorer run. The `REPARERAT:` lines are written into the runs
+   row's `--note`, never pasted into the chat; the BLUF reply mentions a repair in plain
+   words only when it matters to the reviewer (`## Reply wording`). A runs row takes ONE
+   `--note` (`add-run` reads only the first): when step 6 adds comparison rows to the same
+   row, pass both in one value joined with `; `, each repair line prefixed `reparerat: `
+   (the form `korning.mjs` writes).
    `impact.md` is ALWAYS sent to contract-check, never only the spec:
    the script reads the label on `**Deviation from the spec:**` and rejects every
    outcome other than TILLBAKA_TILL_SPEC when the analysis carries `contradiction`
@@ -727,11 +835,13 @@ drift guard.
    has still reached this point has source material the script must get to see). The same script requires
    every `bracket.steps[i]` to be exactly `{anchor, relation, grund}`.
    BROTT that remain after the repairs → redispatch ONCE per
-   the redispatch rule, with the error lines in the prompt. If breaches remain: show them and
-   stop without archiving the reply as a valid round, but do not throw it away:
+   the redispatch rule, with the error lines in the prompt. If breaches remain: stop
+   without archiving the reply as a valid round, but do not throw it away:
    archive it as `score-pricing-vN-avvisad.json` (outside the N count) with
    the check's error lines in a runs row, so even failed runs have a
-   receipt. An `AVVIKELSE:` line is always shown to the user.
+   receipt. Tell the user in plain words that the scoring failed an automatic check
+   twice and was not counted (`## Reply wording`). An `AVVIKELSE:` line is always
+   reported to the user, as what the scorer flagged, in plain words (`## Reply wording`).
 6. **Double run (drift guard).** Run 2 ALWAYS starts at the same time as
    run 1, with the same prompt and in a new isolated run; in production
    both Agent calls are dispatched at once (#500, ADR-0113), the prompts are
@@ -765,13 +875,18 @@ drift guard.
      rows `double_run: "olika"` with `superseded_by: null`).
      Status `assessed`. No customer text may go to
      request-sync while the hold stands (`archive.mjs sparr` is the interpreter).
-     Show in the BLUF reply the evidence behind the disagreement, side by side:
-     the diff rows (outcome, score, anchor, confidence) have no
-     row link to the evidence, `drivers[]` is a list of its own, so show
-     all `drivers[].citat`/`pastaende` from both runs
-     as support for the whole outcome. It was the evidence that decided the ruling in
-     every measured genuine disagreement. Then refer to `döm <item>` and
-     STOP: no rerun, no escalation, no choice of run.
+     Describe in the BLUF reply the evidence behind the disagreement, per
+     `## Reply wording`: first one sentence on what the runs disagree on (the price,
+     the conditions, what is priced, how the work compares with the reference
+     assignments; two runs can disagree at the same score), then for each run, one or
+     two sentences on where it landed and
+     what it leaned on, built from all of its `drivers[].citat`/`pastaende` (the
+     diff rows have no row link to the evidence, `drivers[]` is a list of its own),
+     quoting the source text where it decides. It was the evidence that decided the
+     ruling in every measured genuine disagreement. Never paste the diff rows:
+     the diff rows go into run 1's runs row `--note`, so `detaljer <item> state` shows
+     them. Then refer to `döm <item>` and STOP: no rerun, no escalation, no choice
+     of run.
    - **LIKA_TOLERANS** (a score difference ≤ 0.5 with the same outcome, regardless of
      confidence, a confidence difference alone never rejects a pair within
      the tolerance) → the pair counts as confirmed. The run the verdict line points out
@@ -782,7 +897,12 @@ drift guard.
      the difference lies within the tolerance. Set `double_run: "tolerans"` on
      both runs rows. The tolerance is a deliberate pricing rule: the highest
      value protects the margin, and an average is still never
-     computed. Show the tolerance rows in the BLUF reply. A pair of two
+     computed. Write the tolerance rows into the main file's runs row `--note`, so
+     `detaljer <item> state` shows them, and say in the BLUF reply, per
+     `## Reply wording`, that the two scorings landed close and which score applies;
+     a row that shows something only one run found (an alternative, a different type)
+     is told in a sentence too.
+     A pair of two
      `OVER_SKALAN` with the SAME floor where only the confidence differs
      also writes `LIKA_TOLERANS`, but with no score to compare (OVER_SKALAN carries
      `score: null`) run 1 wins deterministically, never "the highest
@@ -822,16 +942,34 @@ ref-hash.mjs på request-score/references/poangskala.md>
    as it is (`| post | poäng | utfall | konfidens | anm |`, one row per
    directory and part; the score from the files, stop outcomes from `stopp`, OLIKA and
    tolerance in `anm`), NEVER handwritten: a handwritten table
-   misreported in a live run. Under the table, one line
-   per deviation worth seeing (ESKALERA, DELA_UPP,
-   INPUT_SAKNAS, AVVIKELSE lines). The JSON is not shown in the chat:
-   `detaljer <item>` does that. Writing back to the board happens via request-sync
+   misreported in a live run. When the item's design is a draft (the first line of
+   `archive.mjs antaganden <katalog>` reads `Designläge: draft`), paste
+   `node <skillkatalog>/scripts/archive.mjs antaganden <katalog>` under the BLUF as it is: every
+   question the machine answered with the option taken and the options not taken, the review
+   verdict and the spec conflicts; this is the human's reading before `synka`, never
+   handwritten. Under the table, one plain sentence per deviation
+   worth seeing (an escalation, a split, missing input, something the scorer
+   flagged), per `## Reply wording`. The JSON is not shown in the chat:
+   `detaljer <item>` does that. When `bedöm` was entered from `designa` (designa step 6), the BLUF also carries the design doc's path and the commit offer `designa` would have printed. Writing back to the board happens via request-sync
    in the main conversation, only on explicit request, and it is that move
    that takes the item ✅ Spec ok → ⚖️ Poängsättning redo för granskning per
-   the sync's rules. **Close per `## Closing: where the item stands`**
+   the sync's rules. Then record the write: `node <skillkatalog>/scripts/archive.mjs synced <katalog> --offer-status "<the status request-sync set>"` (not on a split part, whose receipt is `set-split-gid`; on the split root, in the run that moves the parent). **Close per `## Closing: where the item stands`**
    (the table from `archive.mjs nasta-steg`, the board line, the ISÄR check, and
    a stop outcome or OLIKA as a choice the user must make; a `DELA_UPP` as the two
-   separate choices in the routing below), never without.
+   separate choices in the routing below; the wrap-up lines at a resting point), never without.
+
+**Several items in one `bedöm`.** `bedöm` with several items, or with none, is a batch:
+with none, the candidates are every item at ✅ Spec ok in request-sync's queue reading
+(as `status`) that has an archive directory. A split root is left out: its parts are priced
+one by one (`<gid>/del-N`) and the parent moves when the last part is in (the routing below).
+Steps 1 and 2 run per item in the main
+conversation (one request-sync read and one `approve` each; an item that fails its gate
+or form check is listed with the reason and left out), then ONE `bedom.mjs` with one
+`--post` per item that passed (two at a time, `--parallellt 2`), then step 8 per item in
+one reply, each with its own `bluf` and `nasta-steg` tables and the starter's USD column.
+The ⚖️ sync still waits for `synka`, the user's explicit request to publish the score (step
+8, in any wording; ADR-0155): after a batch, that request publishes every `priced` item in turn, one request-sync write and one `synced` receipt at a time, never
+in parallel.
 
 **Stop outcomes are routed, never interpreted.** A stop outcome is archived as the round's
 file (step 7), the status becomes `assessed`, never `priced`, and the item is routed:
@@ -839,9 +977,14 @@ file (step 7), the status becomes `assessed`, never `priced`, and the item is ro
 - `DELA_UPP` → the Closing offers two separate yes/no choices (ADR-0144), each answered on its own:
   1. **Split the item** (as before): a new analysis round per part; the parts become
      directories of their own, `del-1/ … del-M/` under the parent's directory, with `split`
-     set (`parent`/`part`/`of`), each `state.json` written by the orchestrator as before
-     (`ny-del` stays the experiment's command: its skeleton carries `experiment: true`,
-     which `set-status priced` refuses), and a part without a ceiling becomes `OVER_SKALAN`,
+     set (`parent`/`part`/`of`), each created with
+     `node <skillkatalog>/scripts/archive.mjs ny-del <förälderns katalog> --typ split --part N --of M`,
+     never a hand-written `state.json`: under a real item at the scorer's `DELA_UPP` it
+     writes a production part (no `experiment` marker, `split.parent` the parent's gid or,
+     without one, its directory name, the parent's `approval` inherited, `part`/`of`
+     integers checked against the siblings under the parent's lock; `--request-id` is
+     optional there and must match the parent) and refuses anything else, exit 2; under
+     an experiment directory it stays the experiment's command. A part without a ceiling becomes `OVER_SKALAN`,
      never DELA_UPP again. Sub-items inherit, per the skill, the item's spec and never write
      one of their own: the parent's `spec.md`, `granskarblock.md` and `design.md` are COPIED
      unchanged into each part directory, separately (one reviewer block per part directory,
@@ -927,14 +1070,20 @@ get wrong:
 - **`underkänn <nr|gid> <reason>`:** the reason is mandatory: without a reason, ask
   for it. request-sync adds the reason as a comment on the item, raises
   the iteration counter and moves the status backward per its backward-move rules.
+  When the move was from ⚖️, request-sync reports it done and the item has an archive
+  directory, record the move: `node <skillkatalog>/scripts/archive.mjs synced <katalog> --offer-status "<the status request-sync set>"`
+  (not on a split part, whose receipt is `split.gid`), so `nasta-steg` stops calling the
+  ⚖️ sync done. A refused move records nothing. `godkänn` records no receipt, and `byt titel` moves no status, so neither writes one.
 - **An item without a board** (archive only): `godkänn` notes
-  `approval: {source: "manual", date}` in `state.json` after an explicit
+  `approval: {source: "manual", by, date}` in `state.json` through
+  `archive.mjs approve <katalog> --source manual --by <name>` after an explicit
   confirmation question. Say clearly that this is not visible to anyone other than the archive.
 - The verbs touch Offer status and comments, never Score, Class or
   description fields.
 - **Close per `## Closing: where the item stands`** when the item has an
   archive directory: the table from `archive.mjs nasta-steg` plus the board line, so
-  the move is visible together with what it opens up for (`designa`, customer, nothing).
+  the move is visible together with what it opens up for (`designa`, customer, nothing),
+  and the wrap-up lines at a resting point.
 
 ## Task: `döm`
 
@@ -957,9 +1106,11 @@ run again for the same item, whose resumability rule re-evaluates a
 `"väntar på dom"` variant and makes the outcome visible to the report when it runs
 (see `utvärdera`).
 
-1. **Show the source material.** Run the comparison script on the current pair and show
-   every diff row with the evidence side by side (as in bedöm step 6). No
-   summary instead of the evidence: the ruling is made on it.
+1. **Show the source material.** Run the comparison script on the current pair and
+   describe every difference it finds, with each run's evidence in sentences and the
+   source quotes that drove it (as in bedöm step 6, `## Reply wording`); the raw rows
+   stay behind `detaljer <item> state`. No summary instead of the evidence: the ruling
+   is made on it.
 2. **Ask the reviewer** (AskUserQuestion) with three paths.
    - **"Run N holds"** (N is 1 or 2, whichever of the two measured
      runs was right): write the ruling via `archive.mjs utslag <katalog> --typ viktning --beslut "<ordagrant>" --beslutad-av <namn> --axel score --n <n> --galler-fil <fil>`.
@@ -1036,13 +1187,21 @@ facts in the analysis above.
 
 ## Task: `detaljer`
 
-`detaljer <gid|slug|nr> [step]`: print archived files for the item: without
+`detaljer <gid|slug|nr> [del-N] [step]`: print archived files for the item: without
 a step, a list of what exists; with a step (`spec`, `granskarblock`, `design`,
-`analys`, `poäng`, `state`) the file's content raw (`design` prints `design.md`);
+`granskning`, `analys`, `poäng`, `state`) the file's content raw (`design` prints `design.md`,
+`granskning` prints `design-review.md` when present);
 `klass` reads older archives'
 classification files (before ADR-0098). For `klass`/`poäng`: the latest
 `-pricing-vN.json` without a suffix (never `-run2`/`-avvisad`).
 
+On a split root, the listing without a step also lists the parts (the `del-1/ … del-M/`
+directories with `split` set), one line each with the part's status and, when recorded,
+its subtask gid (`split.gid`). `detaljer <item> del-N [step]` reads that part's directory
+(the form `<gid>/del-N` that `bedöm` and `bedom.mjs` print is the same address) the same way: its own `impact.md`, pricing files and `state.json`, and the copies of the
+parent's `spec.md`, `granskarblock.md` and `design.md` it inherited. A `del-N` that does
+not exist: say so and list the parts that do.
+
 ## Task: `byt titel`
 
 `byt titel <nr|gid>`: rename the item's Asana task to the spec title, on request only
@@ -1068,9 +1227,9 @@ reaches the name.
 ## Task: `utvärdera`
 
 Model evaluation: `utvärdera <item…> <model…>`, with at least one model. Runs an
-item through the whole flow once per named model, each round with all
-three agents (the specifier, the assessor, the scorer) forced to the round's
-model, so that each round measures an actual model difference instead of noise between
+item through the whole flow once per named model, each round with
+every agent step forced to the round's model (six with `design_mode=draft`: specifier,
+designer, reviewer, assessor, two scorers; `--modell default` keeps the pins), so that each round measures an actual model difference instead of noise between
 a fixed production mix and a single experiment model.
 `default` as the model value gives the agent defaults (today's production mix) and
 runs without an override; the directory name then becomes `baseline/`, for all other values
@@ -1154,6 +1313,11 @@ The question is therefore about the cost, not the risk of losing data.
    carries the same value is a no-op. `sprak` is read from the request's body text by
    the main conversation, just as in `speca` step 2; `korning.mjs` refuses
    to start without it (it never guesses the language, ADR-0114).
+   `design_mode` is optional: `--satt design_mode=draft|intervju` in a call of its own,
+   absent is `intervju`. With `draft`, `korning.mjs` drafts and reviews the design in every
+   part directory that has no `design.md` before the assessor, the same two steps as
+   production (ADR-0156), so a measurement sweep compares machine drafts with human designs
+   on the same specs before any board opts in; a pinned design (step 2) is never redrafted.
 
    At the same time, note the item's `permalink_url` (`--satt permalink_url=<url>`)
    when the gid reply carries it. It comes from the same read-only fetch
@@ -1215,12 +1379,16 @@ The question is therefore about the cost, not the risk of losing data.
    model:
 
    ```
-   node <skillkatalog>/scripts/svep.mjs --repo-rot <abs> --konto <namn> \
+   node <skillkatalog>/scripts/svep.mjs --repo-rot <abs> --arkiv <arkivrot> --konto <namn> \
      --modell <modell> [--modell <modell>]... \
      --exp experiments/<slug> [--exp experiments/<slug>]... \
      [--tickets-parallellt 2] [--modeller-parallellt 4] [--om | --hoppa-över]
    ```
 
+   `--repo-rot <abs>` is the checkout whose code is read; `--arkiv <abs>` is
+   `ARKIVROT:` from `rot` (ADR-0154, ADR-0155); a worktree as `--repo-rot` is
+   fine, the archive root never comes from it.
+
    Every round is started by `svep.mjs` as ONE `korning.mjs` process, and that is
    what gives the measurement for free. The parallel rule: two tickets at a time, at most
    four models for one ticket in flight (`--modeller-parallellt`, #526), and
@@ -1269,7 +1437,8 @@ The question is therefore about the cost, not the risk of losing data.
      a run without the field as measured under unknown concurrency (#526).
    - **The provenance is stamped AFTER `korning.json`, never before**, otherwise
      the stamp lacks the billing data. The command itself computes `skill_sha256`
-     (over `SKILL.md` plus the four agent definitions) and `underlag_sha256`,
+     (over `SKILL.md` plus the six agent definitions, `request-designer` and
+     `request-design-reviewer` among them) and `underlag_sha256`,
      reads `korning.json` and notes `fakturerade_modeller` (the
      canonicalized key set from `modelUsage`, suffixes like `[1m]` stripped,
      ADR-0099), and also stamps every `del-N/`. Without that stamp it is
@@ -1314,8 +1483,8 @@ The question is therefore about the cost, not the risk of losing data.
    anchor resolution, double run per bedöm) against the directory it is given, without
    the approval gate, and touches nothing outside it. If the model is `default`,
    the ordinary prompt templates and the agents' defaults are used, no override. If
-   it is a model name, every agent process gets a model override to it on ALL
-   three agents; the override is `--model` on the process, never a prompt line, and
+   it is a model name, every agent process gets a model override to it on every
+   agent step; the override is `--model` on the process, never a prompt line, and
    the agent definitions are never touched; the effort is taken from the definition
    as usual, except when `korning.mjs` has been given `--effort`: then every
    agent process gets the same `--effort` override, the step entry and `runs[]` carry it
@@ -1504,13 +1673,45 @@ across items with different model sets, or state which outcome
 was correct. The judgment of what is an acceptable outcome is
 the human's, made on experience of the system. The report shows the spread.
 
+## Reply wording: what the reviewer reads
+
+The sentences the orchestrator writes around the scripts' tables (the BLUF lines, how
+two scorings compared, a repair, a retake, a disagreement) say what the result means
+for the reviewer's decision, in plain sentences. The tables from `bluf` and
+`nasta-steg` are pasted as they are; this rule governs the sentences around them.
+
+- **No internal vocabulary in those sentences:** no reference-scale position names
+  (the anchor ids), no rule or protocol terms (floor, ceiling, between, LIKA,
+  LIKA_TOLERANS, OLIKA, BROTT, REPARERAT, contract, tolerance rows) and no archive file
+  names. Outcome codes (ESKALERA, DELA_UPP, INPUT_SAKNAS, OVER_SKALAN,
+  TILLBAKA_TILL_SPEC, KONCEPT, DIAGNOS) are said as what they mean (an escalation, a
+  split, missing input, above the scale and quoted separately). Quoting the item's own
+  source text from `drivers[].citat` is fine.
+- **After a pricing check, answer in this order:**
+  1. Did the two scorings agree, and how close they were.
+  2. Which score applies, with the reason in a clause; when both runs land above the
+     scale, that the item has no score and is quoted separately.
+  3. Anything that should lower trust in the number: a scoring that failed an
+     automatic check and was redone, a repair to customer text, something the scorer
+     flagged, an alternative or a type only one run found. A routine repair that changes nothing the reviewer decides on needs no
+     mention.
+  4. On disagreement, the evidence in sentences: first what the runs disagree on, then
+     per run where it landed and what it leaned on, then the ruling command.
+- **Raw details stay behind `/request detaljer <item>`:** `poäng` prints the
+  scoring, `state` the runs rows with their notes (the repairs, and the comparison rows
+  a `bedöm` double run wrote).
+  Point there in one line when the reviewer may want them.
+
+Example: "Two independent scorings landed close (2.75 and 3), so 3 applies. One of
+them failed an automatic check and was redone once; the redo agreed with the other run."
+
 ## Closing: where the item stands
 
 The tasks `speca`, `designa`, `bedöm`, `godkänn`, `underkänn`, `döm` and `byt titel` ALWAYS end
-with this block (#530). A reply that ends without it leaves
+with this block (#530). A `designa` that continues into `bedöm` (designa step 6) ends the turn with `bedöm`'s block, once: the block is per turn, never per task. A reply that ends without it leaves
 the user to guess where the item stands, what they should do and which command
 runs on, and the archive and the board can then drift apart silently (`spec_draft`
-locally, empty Offer status on the board). The block consists of four parts, in order:
+locally, empty Offer status on the board). The block consists of five parts, in order:
 
 1. **The table from `node <skillkatalog>/scripts/archive.mjs nasta-steg <katalog>`**
    is pasted as it is (`| post | står | väntar på | nästa |`, one row per
@@ -1526,7 +1727,9 @@ locally, empty Offer status on the board). The block consists of four parts, in
 2. **The board line.** If the item has a gid: fetch Offer status via request-sync in
    the main conversation (never from memory) and write the line
    `Bräda: <Offer status>` directly under the table. If the item lacks a gid: write
-   `Bräda: ingen (posten ligger inte på brädan)`. For an item at
+   `Bräda: ingen (posten ligger inte på brädan)`. When step 7 placed the item on the board
+   under `auto_sync_spec`, the board line ends with `(synkad automatiskt, auto_sync_spec)`.
+   For an item at
    ⚖️ Poängsättning redo för granskning, add one line: the analysis and the pricing are
    not on the board (ADR-0144), `detaljer <item>` shows them.
 3. **The ISÄR check.** The archive status and the Offer status must correspond to each other
@@ -1556,22 +1759,49 @@ locally, empty Offer status on the board). The block consists of four parts, in
    are presented as a choice the user must make, not just as routing:
    "here is your choice", followed by the alternatives (for TILLBAKA_TILL_SPEC the
    open choice from `stopp.orsak` or the source file; for OLIKA the two
-   runs' evidence side by side, see bedöm step 6) and the command that
+   runs' evidence in sentences, see bedöm step 6 and `## Reply wording`) and the command that
    carries out the answer (`speca <item>` after `## Avgjorda val` + `respec`,
    `döm <item>`, customer dialogue). A `TILLBAKA_TILL_SPEC` from `impact.md` whose
    `orsak` starts with `design conflict:` routes to `designa <item>` (a redesign, bedöm
-   step 3), not to `speca`; the command for that case is `designa <item>`. A stop that
+   step 3), not to `speca`; the command for that case is `designa <item>`. A row
+   `designed, granskningen stoppade` is a choice: read `design-review.md` (`detaljer <item>`)
+   and run `designa <item>` for the interview, or accept the draft as it stands with
+   `node <skillkatalog>/scripts/archive.mjs add-design-review <katalog> --file <katalog>/design-review.md --verdict fortsätt --model människa`
+   (the human's verdict is a receipt; the reviewer is never run again on the same doc). A
+   `stopp ESKALERA ur design-intervju-kravs-<datum>.md` row is the designer asking for the
+   interview: run `designa <item>`. A stop that
    is only routed is read as "the machine is working on it", and nobody does anything.
    A `DELA_UPP` is presented as two separate yes/no choices, split the item and rewrite
    the parent as an umbrella (bedöm's routing, ADR-0144), each answered on its own.
+   A `pågår` row from `nasta-steg` is not a choice: the item is being priced by
+   `bedom.mjs`, wait for its table.
+
+5. **Wrapping up, at a resting point.** A resting point is a turn that leaves the item
+   on the board at 🔍 (spec review), ⚖️ (price review) or 📤 (done, with the customer),
+   with no part-4 choice open. There, end with a few plain lines on
+   what to do before the session closes, per `## Reply wording` (no anchor ids, protocol
+   terms or archive file names):
+   - Commit the design document and the setup files (the board connection and what
+     `haus apply` wrote), if they are not committed yet; in a temporary workspace, get
+     them onto the main branch before the workspace is removed. This line replaces the
+     separate design commit offer in the same turn.
+   - Keep the specs, analyses and scores out of the customer's history: they stay in the
+     local request archive, which is kept out of git on purpose.
+   - When the customer orders, planning starts from the saved design document; if the
+     spec changes before that, the design is redone first.
+
+   Leave out a line that does not apply (nothing to commit, no design yet). No line about
+   moving the archive out of a workspace: the archive always lives in the main checkout
+   (ADR-0154). An old archive left in a workspace shows up as `archive.mjs rot`'s
+   VARNING line; pass that line on to the user in plain words when it appears.
 
 ## Mechanical scripts: `scripts/`
 
-Twelve deterministic scripts live in the directory: the five checks below,
+Thirteen deterministic scripts live in the directory: the five checks below,
 `kontroll.mjs` which runs the scorer's check chain in one process,
 `archive.mjs` which owns the archive's directories, holds and `experiment.json`,
-`svep.mjs` which is utvärdera's starter,
-`korning.mjs` which is utvärdera's inner process, `report.mjs` which
+`svep.mjs` which is utvärdera's starter, `bedom.mjs` which is bedöm's starter,
+`korning.mjs` which is utvärdera's inner process and bedöm's in production, `report.mjs` which
 renders the model report (see `utvärdera rapport`), `lib.mjs` with the shared
 building blocks and `ref-hash.mjs` for stable hashing. The checks replace
 model runs where the check is pure form:
@@ -1591,7 +1821,10 @@ reference files whose rules it mirrors and WARNS when a catalog bump has changed
   _Lösning_ (_Solution_) is a BRIST. With `--impact` the label on the analysis's row
   `**Deviation from the spec:**` is read and written as `AVVIKELSE: ingen`,
   `AVVIKELSE: precisering` or `AVVIKELSE: motsägelse`; a row without a
-  label is a BRIST. The orchestrator reads candidate and deviation from those
+  label is a BRIST. When the row carries text after its label, the line
+  `AVVIKELSE-TEXT: <text>` follows right after `AVVIKELSE:`; the orchestrator and
+  `korning.mjs` read the design-conflict text from that line, never from the file
+  (ADR-0155). The orchestrator reads candidate and deviation from those
   lines, never from the files itself. The spec rules live in `lib.mjs` (`specForm`),
   shared with `archive.mjs paraply` (ADR-0144).
   With `--design` (`designa` step 3 and `bedöm` gate 1b): `design.md` must exist,
@@ -1607,6 +1840,16 @@ reference files whose rules it mirrors and WARNS when a catalog bump has changed
   match the analysis row's `underlag_sha256["design.md"]`, the same provenance rule
   as for `spec.md`, except that a missing design hash is a BRIST, not a VARNING:
   the design existed when the analysis should have seen it.
+  `--design` also checks the draft's form (ADR-0156): it prints
+  `DESIGN-LÄGE: draft|intervju` (from `mode:` in the frontmatter); on a draft a missing or
+  malformed `## Assumptions`, `## Removed or renamed` or `## Files touched` is a BRIST, on an
+  interactive doc a VARNING; every symbol under `## Removed or renamed` is grepped
+  (`git grep`, in the repo of the canonical file) outside the files under `## Files touched`
+  and a hit is a BRIST with its `file:line`, while a grep that cannot run is a VARNING; and a
+  `state.design.review` verdict `stopp` is a BRIST. `--design-utkast <abs fil> [--repo-rot <abs>]`
+  checks a draft file without `state.design` (the experiment channel): the same section and
+  symbol checks, the frontmatter against the directory's id and the current `spec.md` hash,
+  `## Spec conflicts` `none`, and the `DESIGN-LÄGE:` line; exit 1 on a BRIST.
 - `scripts/resolve-anchors.mjs <run.json> <spec.md> [impact.md]`: resolves
   evidence anchors (`belagg_ankare`/`ankare`) to verbatim source blocks, runs
   BEFORE contract-check. Exit 0 = everything resolved (or no anchors in the file), 1 =
@@ -1664,20 +1907,55 @@ reference files whose rules it mirrors and WARNS when a catalog bump has changed
   byte-identically, writes `state.design` and sets status `designed`; no runs row;
   refuses a file under `.haus-workflow/`, a `priced` or `guard_run` item and any
   standing stop except a design conflict from `impact.md`, which it clears, putting
-  `impact.md` aside as `impact-superseded-<datum>.md`;
+  `impact.md` aside as `impact-superseded-<datum>.md`; also refuses a draft whose sections
+  fail the form check and a removed symbol still in use outside the touched files, and
+  resets `design.review`, ADR-0156;
   exit 2 on every refusal, under the archive lock),
+  `add-design-review <katalog> --file <abs> --verdict fortsätt|stopp --model <m>` (the one
+  writer of `state.design.review`: copies the reviewer's reply byte-identically to
+  `design-review.md`, an earlier one put aside as `design-review-superseded-<datum>.md`,
+  writes `{ verdict, sha256, date, model }` and prints `OK: design-review <verdict> (<sha256>)`;
+  refuses an item without a design, a verdict outside the two literals, an unreadable file
+  and an experiment directory, exit 2; no runs row; under the archive lock; ADR-0156),
+  `antaganden <katalog>` (read-only, what `bedöm` step 8 pastes before `synka`: the line
+  `Designläge: draft|intervju`, then `Antaganden (N):` with one block per assumption (the
+  question, `vald:`, `ej valda:`, `skäl:`), or `Antaganden: inga`, or
+  `Antaganden: avsnitt saknas`, then `Granskning: <verdict> (<K> fynd)` or
+  `Granskning: ingen`, then `Spec conflicts: none` or the count with the lines; refuses an
+  item without `design.md`, exit 2; ADR-0156),
   `set-split-gid <del-katalog> --gid <gid>` (records a split part's Asana subtask gid in
   `split.gid` under the archive lock; refuses a directory without `split`, a gid that is not
   digits only and a different gid already recorded, while the same gid again is a no-op;
   ADR-0144),
+  `approve <katalog> --source asana|manual` (`--status "<Offer status as read>"` for
+  asana, `--by <name>` for manual; the one writer of `state.approval`, under the archive
+  lock, overwriting an earlier note; refuses an asana status whose alphabetic core is not
+  `specok`, manual without `--by` and an experiment directory, exit 2; ADR-0153),
+  `synced <katalog> --offer-status "<Offer status>"` (the one writer of `state.board`, the
+  receipt of a board write: `{ offer_status, date }`, the value verbatim, under the archive
+  lock, overwriting the previous receipt; refuses a missing or empty `--offer-status`, an
+  experiment directory and a split part, whose receipt is `split.gid`, exit 2; `nasta-steg`
+  reads it; ADR-0153),
+  `korning-start <katalog> --pid <N> [--katalog-korning <abs>]` /
+  `korning-slut <katalog>` (the running receipt's two writers, `state.korning`, under the
+  archive lock; start refuses a living pid, exit 2 `pågår (pid N)`, and takes a dead one
+  over; both refuse an experiment directory, exit 2; `bedom.mjs` runs them around every
+  item; ADR-0155),
+  `rot [<dir>]` (takes no archive directory: prints `ARKIVROT: <abs>`, the main checkout's
+  `.haus-workflow/requests/`, and `IGNORERAD: ja|nej`, exit 0, plus a third line
+  `VARNING: arkiv i worktree: <abs> (flytta per ADR-0154)` when the cwd's own checkout holds
+  an archive of its own; refuses a directory outside a git repository, a bare repository, a
+  submodule checkout and a `--separate-git-dir` clone, exit 2; ADR-0154),
   `paraply <katalog> --file <spec>` (rewrites the parent's spec as an umbrella: refuses a
   split part, an item without `approval`, a file that fails the spec form check
   `precheck.mjs` runs, a file without a title line and an unreadable file; puts `spec.md`
   aside as `spec-superseded-<datum>.md`, copies the file in, sets `title` from its heading
   and records `state.umbrella = { date, sha256 }`; approval, design, runs, stops and the
   parts are untouched; prints `OK: paraply <sha256>`; ADR-0144),
-  `stopp <katalog> --utfall KONCEPT|DIAGNOS|TILLBAKA_TILL_SPEC|ESKALERA --kalla spec.md|impact.md|granskarblock.md|design-avvisad-<datum>.md [--orsak <text>]`
+  `stopp <katalog> --utfall KONCEPT|DIAGNOS|TILLBAKA_TILL_SPEC|ESKALERA --kalla spec.md|impact.md|granskarblock.md|design-avvisad-<datum>.md|design-intervju-kravs-<datum>.md [--orsak <text>]`
   (writes `state.stopp` after checking that the source carries the outcome;
+  `--kalla design-intervju-kravs-<datum>.md` requires a first line `INTERVJU KRÄVS: <reason>`
+  and takes only `ESKALERA` with `--orsak`, written by `korning.mjs` in draft mode;
   `--kalla design-avvisad-<datum>.md` requires at least one conflict under
   `## Spec conflicts` in that receipt and takes `TILLBAKA_TILL_SPEC` or `ESKALERA`,
   and replaces a standing design-conflict stop from `impact.md` (any other standing
@@ -1685,8 +1963,9 @@ reference files whose rules it mirrors and WARNS when a catalog bump has changed
   `--kalla impact.md` sets status `assessed`, the other sources do not touch
   the status), `respec <katalog>` (puts `impact.md` aside as
   `impact-superseded-<datum>.md` and `design.md` as `design-superseded-<datum>.md`,
-  resets status (to `spec_draft`), `approval`, `granskarutslag`, `stopp` and
-  `design`, never touches `runs[]`) and
+  resets status (to `spec_draft`), `approval`, `granskarutslag`, `stopp`,
+  `design` and `board`, never touches `runs[]`; `design.review` goes with `design`, and `design-review.md` is put aside as
+  `design-review-superseded-<datum>.md`) and
   `bluf <katalog>` (the markdown table `| post | poäng | utfall | konfidens | anm |`
   over the directory and its `del-N/` recursively, the bundle root as
   `bunt (N delar)`, a split root with `K av M delar prissatta` in `anm`; the values from
@@ -1698,11 +1977,19 @@ reference files whose rules it mirrors and WARNS when a catalog bump has changed
   outcome per the rule table in the script; on a split root the parent row counts the
   parts (`se delarna (K av M delar prissatta)`) until every part is `priced` or
   `OVER_SKALAN` and then names the sync that moves the parent, and a part row says
-  whether its subtask is recorded (`split.gid`); an unknown state is written as
+  whether its subtask is recorded (`split.gid`); a living `korning.pid` gives `pågår (pid N)`
+  with `vänta; bedöm igen när bedom.mjs är klar`, a dead one prefixes the row with
+  `avbruten körning, ` (ADR-0155); a `designed` item whose review stopped gives
+  `| <post> | designed, granskningen stoppade | människa | designa <id> (intervju; se design-review.md) |`
+  (ADR-0156) and an `INTERVJU KRÄVS` stop gives
+  `| <post> | stopp ESKALERA ur design-intervju-kravs-<datum>.md | människa | designa <id> (intervju; se design-intervju-kravs-<datum>.md) |`;
+  an unknown state is written as
   `okänt läge: <status>` with exit 0, a broken `state.json` is exit 2, see
   `## Closing: where the item stands`). `add-run` reads
   `effort:` from the agent definition when `--effort` is omitted and canonicalizes
-  `--model`; `set-status priced` refuses when `stopp` is set.
+  `--model`, and `add-run … [--repo-sha <sha>]` writes `repo_sha` on the row (40 lowercase
+  hex digits, otherwise exit 2; `korning.mjs` passes it on `impact.md` rows, ADR-0155);
+  `set-status priced` refuses when `stopp` is set.
   `experiment <experimentkatalog> [--satt nyckel=värde]... [--variant <modell>]... [--korning <modell>/k<N> --status <klar|ofärdig|stopp|väntar på dom> [--orsak <text>] [--pid <heltal>]]`
   is the ONLY writer of `experiment.json` (see `utvärdera` step 1): a locked
   read-modify-write with an atomic write; the directory and file are created if they are
@@ -1714,7 +2001,7 @@ reference files whose rules it mirrors and WARNS when a catalog bump has changed
   the starter's pid on the row; a status write without `--pid` replaces the row
   entirely, so the pid disappears with the final status (#516). Exit 0 ok, 2
   a usage or state error.
-- `scripts/svep.mjs --repo-rot <abs> --konto <namn> --modell <m>... --exp experiments/<slug>... [--tickets-parallellt 2] [--modeller-parallellt 4] [--om | --hoppa-över]`:
+- `scripts/svep.mjs --repo-rot <abs> [--arkiv <abs>] --konto <namn> --modell <m>... --exp experiments/<slug>... [--tickets-parallellt 2] [--modeller-parallellt 4] [--om | --hoppa-över]`:
   utvärdera's starter (see `utvärdera` step 4 and Batch). Runs two tickets at
   a time with at most four models per ticket in flight, the next ticket when one is completely
   done, and does per round `ny-korning`, the start of `korning.mjs`,
@@ -1725,12 +2012,39 @@ reference files whose rules it mirrors and WARNS when a catalog bump has changed
   `väntar på dom` is tested with `sparr`, `ofärdig` is rerun, except when the row carries
   a living `pid` (in progress in another sweep, skipped as `pågår`, #516). An `--exp` without
   `onskemal.md`, `kallmanifest.md` or `experiment.json` with `request_id` is marked
-  and skipped. Exit 0 when no run became `ofärdig`, 1 otherwise, 2 on a
-  usage error.
-- `scripts/korning.mjs --katalog <abs k<N>> --exp <abs experimentkatalog> --repo-rot <abs> --request-id <gid eller slug> --modell <modell|default> [--claude <kommando>] [--delar-parallellt 2] [--effort <low|medium|high|xhigh|max|default>]`:
+  and skipped. `--arkiv` is the archive root; omitted, it is derived from
+  `--repo-rot`'s git common dir (ADR-0154). Exit 0 when no run became
+  `ofärdig`, 1 otherwise, 2 on a usage error.
+- `scripts/bedom.mjs --repo-rot <abs> [--arkiv <abs>] --post <gid|gid/del-N>... [--parallellt 2] [--max-usd <n>] [--claude <kommando>] [--om] [--sprak <svenska|engelska>]`:
+  production's starter (see `bedöm`), one `korning.mjs --post` process per item, two at a
+  time. `--post` is relative to the archive root (`--arkiv`, or derived from `--repo-rot`'s
+  git common dir as in `svep.mjs`); an absolute `--post` is a usage error. Per item: the
+  receipt via `archive.mjs korning-start`/`korning-slut` (the latter always, also when the
+  run fails), the run directory `<post>/korningar/<stämpel>/` with `stderr.log`,
+  `matning.json` and the raw replies under `steg/`, the cap from `request.json`
+  (`max_usd_per_batch`, default 100 USD, `--max-usd` overrides; once the finished items'
+  USD, plus each item in flight at the average of the finished ones, reaches it no further
+  item starts, `tak nått`, and a running item is never killed), `--startare <its own pid>` on
+  every `korning.mjs` it starts, SIGINT and SIGTERM both killing the running items and
+  clearing their receipts (exit 130),
+  and the final table `| post | status | poäng | utfall | wall | USD |` in argument order
+  with the line `tak: <n> USD, använt: <m> USD`. Exit 0 when every item ended in a result
+  (`priced`, `stopp`, `väntar på dom`), 1 on `ofärdig`, `pågår`, `grind` or `tak nått`, 2
+  on a usage error (ADR-0155).
+  `bedom.mjs` takes `--design-mode draft|intervju` and passes it to every `korning.mjs` it
+  starts; without the flag the value is `request.json`'s `design_mode` (`interview` is read
+  as `intervju`), and without that key `intervju` (ADR-0156).
+- `scripts/korning.mjs --katalog <abs k<N>> --exp <abs experimentkatalog> --repo-rot <abs> --request-id <gid eller slug> --modell <modell|default> [--claude <kommando>] [--delar-parallellt 2] [--effort <low|medium|high|xhigh|max|default>] [--sprak <svenska|engelska>]`:
   utvärdera's inner process (see `utvärdera` step 4), a script without
-  model turns. Builds the prompts from `onskemal.md`, `kallmanifest.md` and
-  `experiment.json`, where `sprak` is MANDATORY: if it is missing, an
+  model turns. In production (`bedom.mjs`) `--post <abs postkatalog>` replaces `--exp`
+  plus `--request-id` (giving `--request-id` with `--post` is a usage error, the id comes
+  from the item's `state.json`) and runs `bedöm` steps 3 to 7 on the item directory,
+  printing `STATUS: <priced|assessed|stopp|väntar på dom|ofärdig>` on stdout (ADR-0155).
+  `--startare <pid>` (production only, a usage error with `--exp`) is the starter's pid,
+  checked every 30 s: a dead starter aborts the run like SIGTERM (exit 130).
+  `--sprak` overrides the language in both modes; in production it otherwise comes from
+  `SPRÅK:` in `kallmanifest.md`, and a missing language is exit 2. In utvärdera it builds
+  the prompts from `onskemal.md`, `kallmanifest.md` and `experiment.json`, where `sprak` is MANDATORY: if it is missing, an
   instruction is written on stderr and the script ends with exit 2 before any prompt
   is built (ADR-0114). Every agent step is a `claude -p --agent` process of its own with
   the raw reply under `k<N>/steg/`, and the checks (precheck, `kontroll.mjs`,
@@ -1742,9 +2056,22 @@ reference files whose rules it mirrors and WARNS when a catalog bump has changed
   before; everything else is passed to the process and to `add-run --effort` as
   requested, never verified (#525). `korning.json` is the sum over the steps
   (`korningsmotor`, `steg` with `effort_begard` and `samtidiga_processer` per
-  step, `modellverifiering` plus the report's usual keys). Exit 0 when
-  the run went through, 1 when at least one part became unfinished, 2 on a
-  usage error.
+  step, `modellverifiering` plus the report's usual keys). The impact row gets
+  `--repo-sha`, the checkout's `HEAD` when the assessor read it (ADR-0155).
+  `korning.mjs` takes `--design-mode draft|intervju` (default `intervju`; in utvärdera the
+  value is `design_mode` in `experiment.json`). With `draft`, a part without `design.md`
+  (in production: and not a split part) gets two steps before the assessor (ADR-0156):
+  `request-designer`, whose doc is written at the `designa` path in `--repo-rot` and recorded
+  with `add-design` (in utvärdera written as the part's `design.md` and checked with
+  `precheck --design-utkast`), then `request-design-reviewer`, whose reply is recorded with
+  `add-design-review` (in utvärdera copied as the part's `design-review.md`).
+  `INTERVJU KRÄVS:` gives `stopp` with `intervju krävs: <reason>` and writes nothing, a
+  refusal on `## Spec conflicts` gives the receipt and the stop of `designa` step 5, any other
+  refusal, BRIST or symbol hit gives `ofärdig`, and `VERDICT: stopp` gives `stopp` with
+  `designgranskning: <first finding>`. A reviewer reply without a `VERDICT:` first line is a
+  script-proven breach and is redispatched ONCE; the verdict itself never is.
+  Exit 0 when the run went through, 1 when at least one part became
+  unfinished, 2 on a usage error.
 - `scripts/kontroll.mjs <svar.txt> <spec.md> <impact.md> --expect-id <id> --ut <ren.json> [--fixa-kundtext] [--claude <kommando>] [--fix-modell sonnet] [--fix-fil <rå.json>]`:
   the scorer's check chain in ONE process (see `bedöm` step 5): tail lines,
   JSON extraction, deterministic repairs, then check-lasta,
@@ -1769,7 +2096,7 @@ an item that passes can still be stopped by the protocols themselves, and that i
 their job. On a script error (exit 2 or a crash): show the error and stop, never
 continue as if the check had passed.
 
-## Dispatch contract (applies to all three agents)
+## Dispatch contract (applies to every agent)
 
 - The prompt is self-contained: the agent does not see this conversation.
 - **Prompt templates, the labels verbatim:**
@@ -1825,6 +2152,58 @@ continue as if the check had passed.
   The umbrella dispatch writes no runs row: `state.umbrella` is its receipt once `paraply`
   accepts the draft.
 
+  `request-designer` (draft mode, ADR-0156):
+
+  ```
+  SPEC:
+  <spec.md bytetroget>
+
+  GRANSKARBLOCK:
+  <granskarblock.md bytetroget>
+
+  URSPRUNGLIG FÖRFRÅGAN:
+  <onskemal.md ordagrant när posten har en, annars "(saknas i arkivet)": då arbetar designern ur specen ensam>
+
+  REPO-ROT: <absolut sökväg>
+  DEPS: <installerade | saknas>
+  SPRÅK: <svenska | engelska>
+  KÄLLOR:
+  <kallmanifest.md ordagrant, eller "inga">
+
+  KONFIG: <absolut sökväg till .haus-workflow/workflow-config.md>
+  ID: <katalogens id>
+  SPEC-SHA256: <ref-hash.mjs på spec.md>
+  DATUM: <YYYY-MM-DD>
+  ```
+
+  `request-design-reviewer` (draft mode, ADR-0156):
+
+  ```
+  SPEC:
+  <spec.md bytetroget>
+
+  GRANSKARBLOCK:
+  <granskarblock.md bytetroget>
+
+  DESIGN:
+  <design.md bytetroget, utkastet som granskas>
+
+  REPO-ROT: <absolut sökväg>
+  DEPS: <installerade | saknas>
+  SPRÅK: <svenska | engelska>
+  ```
+
+  The designer and the reviewer are dispatched only by `korning.mjs`, never from the main
+  conversation: a design without a human is a background step, the interview is `designa`.
+  The designer triages as brainstorming does (spike, bounded, architectural) without
+  loading it, takes the recommended option on every HOW question and records it under
+  `## Assumptions`, puts every WHAT question under `## Spec conflicts` unanswered and every
+  defect outside the spec's _När det är klart_ under `## Out of scope`; it replies with the
+  doc, or with the single line `INTERVJU KRÄVS: <reason>` for an architectural item. The
+  reviewer verifies and never redesigns: its reply starts with `VERDICT: fortsätt` or
+  `VERDICT: stopp`, then `## Findings`. `KONFIG:` is a path the designer reads for repo-wide
+  answers (the test command, the formatter); the file is never pasted.
+
   `request-assessor`:
 
   ```
@@ -1865,6 +2244,11 @@ continue as if the check had passed.
   `designa`), pasted, never pointed to by path: the hard rule "documents are pasted
   byte-faithfully" applies. Without a design (a split part of a parent analysed under
   the old rules, or a standalone assessor run) the block is the single line `DESIGN: ingen`.
+  In draft mode (ADR-0156) the doc is a draft (`mode: draft`) written by `request-designer`,
+  and the reviewer's verdict sits beside it in the archive as `design-review.md`. The block
+  is pasted the same way and the assessor's rule is unchanged: every decision in it is
+  taken, the machine-taken ones included; the human reads those in the Closing
+  (`antaganden`), not in a reopened analysis.
 
   `request-scorer` (source material by path, ADR-0099: never pasted):
 
@@ -1888,12 +2272,12 @@ continue as if the check had passed.
   `contract-check --sprak` rejects customer text in the wrong language.
 
 - **Tail lines after the main content**, in this order: `MODELL: <modell-id>`
-  (all agents), `LÄSTA: <kommaseparerade absoluta sökvägar eller "inga">`
+  (all agents; the designer's `INTERVJU KRÄVS:` reply is the one line and carries none), `LÄSTA: <kommaseparerade absoluta sökvägar eller "inga">`
   (only the scorer: its Read is prompt-restricted, the line makes the rule
   reviewable), `AVVIKELSE: <text>` (only if something in the prompt broke the agent's
   rules, e.g. a price list in the source material). The tail lines are stripped before every
   mechanical check and before archiving; MODELL is noted in the runs row,
-  AVVIKELSE is always shown to the user. LÄSTA is checked MECHANICALLY by
+  AVVIKELSE is always reported to the user, in plain words (`## Reply wording`). LÄSTA is checked MECHANICALLY by
   `scripts/check-lasta.mjs` (ADR-0099): the line must carry the source material's two
   pointed-out paths (both mandatory) and may beyond that carry only files
   under the skill's own directory (the reference file); everything else is a
@@ -1912,7 +2296,9 @@ continue as if the check had passed.
   with an `-avvisad` suffix (pricing: `<typ>-pricing-vN-avvisad.json`;
   text steps: `spec-avvisad-<datum>.md` / `impact-avvisad-<datum>.md`) and gets
   a runs row of its own with the check's error lines: two runs of the same step
-  without a trace of why is exactly the gap the receipt closes.
+  without a trace of why is exactly the gap the receipt closes. In the chat, a retake is
+  told in plain words ("one scoring failed an automatic check and was redone once"),
+  never by its archive file name (`## Reply wording`).
 
 ## Do not use when