capy-ai · Oct 19, 2026 · Oct 18, 2026 · Oct 19, 2026 · Oct 19, 2026 · Oct 19, 2026
diff --git a/BROWSER.md b/BROWSER.md
@@ -584,7 +584,7 @@ SITE=app.example.com
 DOMAIN=example.com
 $B goto "https://$SITE"
 $B cookie-import-browser opera-gx --domain "$DOMAIN" --profile Default   # or: opera
-$B reload                                                               # confirm the intended account
+$B reload                                                               # view the page; only --verify-auth proves sign-in
 ```
 
 Run `$B cookie-import-browser` with no flags to see which browsers were detected. `--profile` can be omitted when only one profile has cookies for the domain. If the receipt reports App-Bound Encryption, run `$B handoff`, sign in to the intended account in the window that opens, then `$B resume` (needs a display).

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,18 @@
 # Changelog
 
+## [1.91.150.0] - 2026-10-19
+
+**The paid evals that still pinned older models now run on the current generation, and two cases that went red on the move are repaired at their cause.** Garry's direction (2026-10-18) is to stop testing older frontier models, so these files are re-baselined as a fresh measured reference rather than held to a threshold change.
+
+### Changed
+- **Old-model pins:** the `claude-opus-4-7` pins in `skill-e2e-plan`, `skill-e2e-plan-prosons`, `skill-e2e-design`, `skill-e2e-office-hours-phase4` and `skill-e2e-qa-bugs` now follow `resolveEvalModel('capture')` (Opus 5.5). The overlay benchmark fixtures pin `claude-opus-5-5` and `claude-sonnet-5-5`, the cookie workflow judge pins `claude-sonnet-5-5`, and `codex-sol-scope-termination` runs on `gpt-6.1-sol` while still rendering the `gpt-5.6-sol` overlay.
+- **/setup-browser-cookies:** the workflow is now five ordered steps. It checks for Aside first and ends there on `READY`, since no import is needed. It then asks every decision (source and profile, domains and target, verification, storage reset, all cookies) before any `$B` command, prepares gstack's browser (verification variables before the daemon starts, the fallback, the `Mode: cdp` stop), imports, and reports on three lines: Imported, Storage reset, Sign-in. The `{{BROWSE_SETUP}}` block, which contradicted the Aside probe, is gone from this skill.
+- **Gate integrity:** no threshold, trial count, case kind or assertion changed. Free-test edits follow the new cookie template: two phrase and order assertions, the approval fixture's reverse edits, and the positional-literals checksum list, which no longer includes this skill.
+
+### Fixed
+- **plan-review-prosons-hardstop-neg:** on Opus 5.5 one trial in three stopped with a provider refusal (`stop_reason: refusal`, category `reasoning_extraction`) on the turn that writes the captured question. Replays of the refused request were stopped 7 of 16 times with the capture prompt's "verbatim" and "Do NOT paraphrase. This is a format-capture test." wording and 0 of 24 without it. The shared prosons capture prompt now asks for "the text of the single AskUserQuestion you would have made", with the same required parts. The decision content and every assertion are unchanged; all four prosons cases pass 3/3.
+- **Cookie workflow judge:** on `claude-sonnet-5-5` it scored clarity 3, completeness 4, actionability 3 against the old template. The rewritten template scores 4/4/4 with unchanged thresholds.
+
 ## [1.91.149.0] - 2026-10-19
 
 **Paid eval cases can now run on the Codex host, and the Monday census runs two arms: Claude Opus 5.5 and Codex gpt-6.1-sol.** `EVALS_HOST=codex` runs the 52 adapter-only gate cases through a full Codex runtime per case. On its first pilot gpt-6.1-sol passed 48 of 52 with no retries, at $4.63 for the 52 cases, against $28.93 for the same cases on Fable (0.16×). Claude-only cases are skipped on Codex with a recorded reason and never count as failed. GSTA-49 TM2.

diff --git a/TODOS.md b/TODOS.md
@@ -1645,19 +1645,24 @@ and `test/dx-selected-navigation-ap.test.ts`. One shared table run once against
 only after `engFirstReviewAUQ` checks native completion once at entry; today each branch gates it
 separately, so the change alters a paid verdict and needs its own paid run.
 
-### P3: Re-pin the five remaining claude-opus-4-7 paid files
-
-**What:** The 2026-09 audit moved seven paid evals to the default capture model (`resolveEvalModel('capture')`).
-`skill-e2e-design`, `skill-e2e-office-hours-phase4`, `skill-e2e-plan-prosons` and `skill-e2e-plan` keep
-`claude-opus-4-7` because six cases failed on the default model in one run (plan-design-review-plan-mode timeout,
-office-hours-phase4-fork format, plan-review-prosons-neutral-neg missing output, plan-ceo-review-selective and
-plan-eng-review 600 s timeouts, plan-ceo-review-expansion-energy posture score 3). plan-review-report, the one PR-gate
-case in `skill-e2e-plan`, left the pin on 2026-10-18 and follows the capture default. `skill-e2e-qa-bugs` returned
-to `claude-opus-4-7` after `qa-b6-static` timed out on the default model in two of three runs (census 36597762183
-and a targeted local rerun): each time the stream stopped mid-message, with no pending tool, right after the model
-found the disabled submit button, and emitted nothing until the 300 s case deadline. They measure an old model.
-
-**Re-entry:** fix the prompt, budget or rubric so each case passes on the default model in one run, then drop the pin.
+### P3: Follow-ups from re-baselining the old-model paid pins on the current generation
+
+**What:** GSTA-49 TM3 (2026-10-18) moved every remaining old-model paid pin to the current generation: the
+`claude-opus-4-7` pins in `skill-e2e-{plan,plan-prosons,design,office-hours-phase4,qa-bugs}` follow
+`resolveEvalModel('capture')`, the overlay fixtures pin `claude-opus-5-5` and `claude-sonnet-5-5`, the cookie workflow
+judge pins `claude-sonnet-5-5`, and `codex-sol-scope-termination` runs on `gpt-6.1-sol` (still rendering the
+`gpt-5.6-sol` overlay). Three things remain:
+
+- The base `claude-dedicated-tools-vs-bash` fixture now duplicates `claude-dedicated-tools-vs-bash-opus-5-5`.
+- The `opus-4-7-*` fixtures measure `model-overlays/opus-4-7.md` on Opus 5.5, which gets the generic `claude`
+  overlay in product.
+- Opus 5.5 refuses (stop category `reasoning_extraction`) capture prompts that ask for the "verbatim text" of a
+  question "you would have made" with "Do NOT paraphrase"; that wording stopped 7 of 16 replays of a prosons
+  request and 0 of 24 without it. `skill-e2e-plan-prosons` dropped it; `skill-e2e-plan-format` and
+  `skill-e2e-office-hours-phase4` still use it.
+
+**Re-entry:** retire one dedicated-tools fixture, decide the opus-4-7 overlay's future, and drop the extraction
+wording from the remaining capture prompts with a paid run each.
 
 ### P3: Retire the unused CEO payment seeder
 

diff --git a/VERSION b/VERSION
@@ -1 +1 @@
-1.91.149.0
+1.91.150.0
diff --git a/agents-digest/gstack-AGENTS.md b/agents-digest/gstack-AGENTS.md
@@ -1,4 +1,4 @@
-# gstack digest v1.91.149.0 — regenerate/re-copy after upgrading gstack
+# gstack digest v1.91.150.0 — regenerate/re-copy after upgrading gstack
 
 Behavioral rules from gstack (https://github.com/garrytan/gstack), compressed
 for agent hosts without a full skill install. The full skills add workflows,

diff --git a/docs/OVERLAY_BENCHMARK_CONTRACT.md b/docs/OVERLAY_BENCHMARK_CONTRACT.md
@@ -14,8 +14,10 @@ Selection IDs for the six retained paid cases stay unchanged.
 ## Blocking checks
 
 Both arms use the real Claude Code preset; ON appends the resolved overlay and
-OFF does not. Each retained fixture runs ten samples per arm, separately for
-`claude-opus-4-7` and `claude-sonnet-4-6`. Each model keeps its own results.
+OFF does not. Each retained fixture runs ten samples per arm on its pinned
+model: `claude-opus-5-5`, plus `claude-sonnet-5-5` for the dedicated-tools `-sonnet` case.
+Each model keeps its own results. The `opus-4-7-*` cases keep their overlay
+(`model-overlays/opus-4-7.md`) but run on Opus 5.5 since GSTA-49.
 
 Every planned sample in both arms must execute successfully, produce exactly
 one successful native terminal result without an error flag, preserve workspace
@@ -38,7 +40,7 @@ are not reinterpreted.
 | `opus-4-7-effort-match-trivial` | Read-only workspace; native final JSON contains the exact version string | Correct answer in every ON sample | Native `usage.output_tokens_details.thinking_tokens` |
 | `opus-4-7-literal-interpretation` | Only `src/auth.ts`, `src/billing.ts`, and `src/notifications.ts` may change; public tests stay frozen; independent behavior checks run outside the writable fixture | All three target behaviors pass in every ON sample | Correct target behaviors, 0..3 |
 
-Each family also has a `-sonnet` case. Output validation checks the native final
+The dedicated-tools family also has `-sonnet` and `-opus-5-5` cases. Output validation checks the native final
 answer, so a correct word mentioned earlier cannot conceal a wrong final answer.
 The read-only prompts explicitly request the JSON shape. These exact-answer
 prompts are new in v2: compare ON/OFF within the same contract and fixture, not

diff --git a/package.json b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "gstack",
-  "version": "1.91.149",
+  "version": "1.91.150",
   "description": "Garry's Stack — Claude Code skills + fast headless browser. One repo, one install, entire AI engineering workflow.",
   "license": "MIT",
   "type": "module",

diff --git a/setup-browser-cookies/SKILL.md b/setup-browser-cookies/SKILL.md
@@ -152,83 +152,59 @@ Skills that run plan reviews (`/plan-*-review`, `/codex review`) include the EXI
 
 # Setup Browser Cookies
 
-## 1. Choose the browser
+## 1. Check for Aside
 
-Use this checkout as the gstack root if it contains `BROWSER.md` and `browse/SKILL.md`; otherwise use the installed root containing `bin/gstack-skill-start`, never a generated host stub. Read that root's `browse/SKILL.md` **BROWSER SETUP** section and run its probe first. On `READY`, stop importing: use Aside's sessions or ask the user to sign in there. Otherwise follow the probe's fallback handling, then continue below.
+Use this checkout as the gstack root if it contains `BROWSER.md` and `browse/SKILL.md`; otherwise use the installed root containing `bin/gstack-skill-start`, never a generated host stub. Read that root's `browse/SKILL.md` **BROWSER SETUP** section and run its Aside probe.
 
-## SETUP (run this check BEFORE any browse command)
+- **Aside probe prints `READY`:** no import is needed, because Aside already has the user's signed-in sessions. Tell the user to use Aside, or to sign in there, and end this skill.
+- **Any other result:** continue with gstack's own headless browser (`$B`).
 
-```bash
-_ROOT=$(git rev-parse --show-toplevel 2>/dev/null)
-B=""
-[ -n "$_ROOT" ] && [ -x "$_ROOT/.claude/skills/gstack/browse/dist/browse" ] && B="$_ROOT/.claude/skills/gstack/browse/dist/browse"
-[ -z "$B" ] && B="$HOME/.claude/skills/gstack/browse/dist/browse"
-if [ -x "$B" ] && "${B%/browse/*}/bin/gstack-browser-ensure"; then
-  echo "READY: $B"
-else
-  echo "NEEDS_SETUP"
-fi
-```
+## 2. Get every decision from the user first
 
-`gstack-browser-ensure` installs Chromium on first use when the install skipped it (lazy browser); it prints `BROWSER_OK` or `BROWSER_UNAVAILABLE <reason>` first.
-
-If `NEEDS_SETUP`:
-1. Tell the user: "gstack browse needs a one-time build (~10 seconds). OK to proceed?" Then STOP and wait.
-2. Run: `cd <SKILL_DIR> && ./setup`
-3. If `bun` is not installed:
-   ```bash
-   if ! command -v bun >/dev/null 2>&1; then
-     BUN_VERSION="1.4.2"
-     BUN_INSTALL_SHA="bab8acfb046aac8c72407bdcce903957665d655d7acaa3e11c7c4616beae68dd"
-     tmpfile=$(mktemp "${TMPDIR:-/tmp}/bun-install.XXXXXX")
-     curl -fsSL "https://bun.sh/install" -o "$tmpfile"
-     # shasum is macOS/perl; coreutils-only Linux ships sha256sum instead —
-     # resolve whichever exists so the verify never fails on a missing tool.
-     if command -v sha256sum >/dev/null 2>&1; then
-       actual_sha=$(sha256sum < "$tmpfile" | awk '{print $(1)}')
-     else
-       actual_sha=$(shasum -a 256 < "$tmpfile" | awk '{print $(1)}')
-     fi
-     if [ "$actual_sha" != "$BUN_INSTALL_SHA" ]; then
-       echo "ERROR: bun install script checksum mismatch" >&2
-       echo "  expected: $BUN_INSTALL_SHA" >&2
-       echo "  got:      $actual_sha" >&2
-       rm "$tmpfile"; exit 1
-     fi
-     bash "$tmpfile" "bun-v$BUN_VERSION"
-     rm "$tmpfile"
-   fi
-   ```
+Ask these together, before running any `$B` command. Never guess an answer or treat the default Comet browser as consent.
 
-```bash
-$B status
-```
-If status says `Mode: cdp`, stop: the real browser already has sessions.
+1. **Source browser and profile.** `--profile` takes the profile's directory (`Default`, `Profile 2`), not its display name. A profile that cannot be read is unknown, not empty.
+2. **Domains to import,** and the **target**: one page on a chosen domain that the session should open, by URL.
+3. **Sign-in verification (`--verify-auth`)?** Default no. It reloads the target and reports **verified** only after a successful same-origin response with exactly one visible element, found by a CSS selector, whose whitespace-normalized text equals the expected identity. If yes, the user supplies the selector and the expected identity privately.
+4. **Clear stale storage (`--clear-storage`)?** Default no; needs explicit approval. Chromium only. It clears localStorage for the target's origin (shared by that origin's tabs) and sessionStorage for the target tab, and nothing else; other browser engines reject the reset but still import. A failed reset may already have cleared some storage and is not rolled back.
+5. **Every cookie in the profile (`--all`)?** Default no; needs explicit consent. It cannot be combined with a domain or with clearing storage.
 
-## 2. Confirm options before import
+Use the picker when any of answers 1-2 is still open; it asks for them. Use direct import when the user has named the browser, the profile directory and the domains.
 
-Open the known target and keep its tab unchanged. Both options default off and require explicit request:
+## 3. Prepare gstack's browser
 
-- **`--verify-auth` / picker checkbox:** reloads the target. Have the user privately configure daemon `GSTACK_COOKIE_AUTH_SELECTOR` and `GSTACK_COOKIE_AUTH_EXPECTED_IDENTITY` **before startup**. Never invent values or assume CLI env reconfigures an existing daemon. Missing config rejects before mutation. Require a successful same-origin response and exactly one visible element whose whitespace-normalized text exactly matches the expected identity.
-- **`--clear-storage`:** for suspected stale storage, obtain explicit approval. Chromium only: clears captured-origin localStorage (shared across same-origin context tabs) and target-tab sessionStorage. Other origins, other tabs' sessionStorage, IndexedDB, and service workers stay intact. It uses an isolated world/native deadline; other engines reject reset, not imports/auth checks. Never auto-approve or claim rollback after partial failure.
+1. If the user wants verification, have them set `GSTACK_COOKIE_AUTH_SELECTOR` and `GSTACK_COOKIE_AUTH_EXPECTED_IDENTITY` privately in the shell that runs `$B`. The browser daemon reads them only when it starts, so if a gstack daemon is already running, run `$B stop` first. Never invent the values.
+2. Follow `browse/SKILL.md`'s **Browser fallback** section to find `$B`, building it if the check prints `NEEDS_SETUP`.
+3. Run `$B status`. If it reports `Mode: cdp`, `$B` is attached to the user's real browser and already has its sessions: tell the user no import is needed and end this skill.
+4. Open the target URL from answer 2 and leave that tab alone until the import finishes:
 
-## 3. Select source and scope
+```bash
+$B goto "<url>"
+```
+
+## 4. Import
+
+Picker (the user confirms the browser, profile and domains there, ticks verification if they asked for it, then tells you when done; the link works once, for five minutes, so rerun the command if it expires):
 
 ```bash
 $B cookie-import-browser
 ```
 
-Ask the user to choose browser, account/profile, and domains, then say when done. Never guess accounts or treat default Comet as consent. Unreadable profiles are unknown, not empty. Rerun for an expired five-minute one-use link.
+Direct import: run `$B cookie-import-browser` with the chosen browser name, one `--domain` per run and `--profile` set to the profile directory (not its display name); add `--verify-auth` or `--clear-storage` only if the user said yes in step 2.
+
+If the import rejects because verification settings are missing, nothing changed: tell the user which variable is missing and return to step 3. Read that same root's `BROWSER.md`, **Choosing a source and checking sign-in**, for other errors and receipt failure reasons; it has the error tables, examples, profile labels, supported sources and platform setup.
 
-Direct import: pass the chosen browser and `--domain` after navigating to a matching target. `--profile` takes a directory, not a display name; omit only for an unambiguous relevant profile, otherwise use the picker. `--all` requires consent for all non-expired profile cookies; it cannot accompany `--domain` or `--clear-storage`.
+## 5. Report honestly
 
-Read that same root's `BROWSER.md`, **Choosing a source and checking sign-in**, for examples, profile labels, supported sources and platform setup.
+Report from the import receipt, not from raw `$B cookies`, as three lines:
 
-## 4. Report honestly
+- `Imported: <count> (<imported|partial|empty|failed>, as the receipt says) from <browser> for <domains>`
+- `Storage reset: <done|failed|not requested>`
+- `Sign-in: <verified|not verified|not checked>`
 
-Report receipt/picker counts, partial/zero/error and reset outcomes, not raw `$B cookies`. Imports affect the context, not one tab. **Not checked** means no requested check; **not verified** means it failed; **verified** requires positive target evidence. Zero imports, counts, or HTTP 200 never prove login.
+**Not checked** means verification was not requested; **not verified** means it ran and failed; **verified** needs the identity match. Cookie counts, zero imports and HTTP 200 never prove sign-in. Imported cookies apply to the whole browser context, not one tab.
 
-Never request/publish cookie values, passwords, identity/profile text, session details, or raw errors in public logs.
+Never request or publish cookie values, passwords, identity or profile text, session details, or raw errors in public logs.
 
 ## Platform boundaries