Add paired main/full content requirements for 530 frozen tasks

README.md

@@ -4,7 +4,7 @@ license: apache-2.0
 tags:
 - benchmark
 - evaluation
-- browser-eval
+- webstone
 - web-fetch
 - web-scraping
 - web-content-extraction
@@ -79,14 +79,14 @@ Short name: **LB-WF**.
 
 LexBench-Web-Fetch evaluates whether a browser, extraction tool or hosted fetch service returns the content an AI agent needs from a URL. Its 550 task entries cover page bodies, structured records with usable links, resource discovery and file delivery. Each task declares what the returned answer must contain, so the same acceptance rules apply to every participant.
 
-Use it to compare live-site fetching, diagnose incomplete extraction, or test a saved response after changing a parser. [Browser Eval](https://github.com/lexmount/browser-eval) provides the scorer, execution adapters, evidence storage and comparison reports. Scoring saved answers requires no browser, provider account or LLM judge.
+Use it to compare live-site fetching, diagnose incomplete extraction, or test a saved response after changing a parser. [Webstone](https://github.com/lexmount/webstone) provides the scorer, execution adapters, evidence storage and comparison reports. Scoring saved answers requires no browser, provider account or LLM judge.
 
 **License.** The `apache-2.0` license covers the project-authored code, documentation and task requirements. The reference texts included for evaluation under `definitions/materials/` are third-party content. Apache-2.0 does not cover them, and they keep their source terms; see [THIRD_PARTY.md](THIRD_PARTY.md).
 
 
 ## Dataset versions
 
-This snapshot is **v2.1.0: 550 tasks**. See [version history and comparison rules](VERSIONING.md) for the 550-task and 430-task editions, version increments and immutable revision selection. The active population is recorded in `manifest.json` under `edition`; retained historical definitions do not increase that population.
+This snapshot is **v2.2.0: 550 tasks**. See [version history and comparison rules](VERSIONING.md) for the 550-task and 430-task editions, version increments and immutable revision selection. The active population is recorded in `manifest.json` under `edition`; retained historical definitions do not increase that population.
 
 ## 530-task revision
 
@@ -133,7 +133,7 @@ After exporting a suite with the compatible exporter and `--revision primary-con
 uv run --frozen python tools/task_population.py --suite webfetch
 ```
 
-With the compatible Browser Eval population support, prefer
+With the compatible Webstone population support, prefer
 `--population /path/to/definitions/populations/primary-content-20261103.json --population-suite webfetch`
 on the certified runner, using a dataset exported with this revision. Repeat for
 each of the five suites. The runner checks the selection against the exported
@@ -175,7 +175,7 @@ A method receives the task URL and prompt. Its response is checked against the t
 | Resources | Returned resource URLs, permitted variants, and target-only output requirements where declared |
 | Files | Delivered bytes, declared file type and integrity requirements |
 
-Contracts can combine these checks. A successful request, nonempty answer or long response alone does not establish task completion. The scorer keeps binary acceptance separate from text-quality diagnostics. Missing reference coverage remains unavailable. Exact requirements and thresholds are in each task; the [metrics guide](https://github.com/lexmount/browser-eval/blob/main/docs/metrics.md) explains their interpretation.
+Contracts can combine these checks. A successful request, nonempty answer or long response alone does not establish task completion. The scorer keeps binary acceptance separate from text-quality diagnostics. Missing reference coverage remains unavailable. Exact requirements and thresholds are in each task; the [metrics guide](https://github.com/lexmount/webstone/blob/main/docs/metrics.md) explains their interpretation.
 
 Web Fetch runs call the tested tools directly. An agent does not choose browser actions during execution. Preserve the requested initial-load scope: clicking, logging in, expanding hidden content or paginating changes the work unless a task explicitly permits it. In the optional article extension, a qualified displayed excerpt is scored as that excerpt, and changing feeds use their own post-collection contracts.
 
@@ -209,7 +209,7 @@ Replace the revision with the full dataset commit you intend to use. Select anot
 | --- | --- |
 | `id` | Task identifier within its source suite |
 | `input` | Source URL |
-| `benchmark_category` | Task category shared with Browser Eval's report grouping |
+| `benchmark_category` | Task category shared with Webstone's report grouping |
 | `source_suite` | Original scorer suite ID, retained independently of the benchmark category |
 | `track` | Original source-group value from the task definition |
 | `category` | Original finer category, preserved exactly as authored |
@@ -223,18 +223,18 @@ The viewer supports inspection and integration. Viewer rows show hashes in place
 
 ## Validate, score and run
 
-From a Browser Eval checkout with [uv](https://docs.astral.sh/uv/getting-started/installation/) installed:
+From a Webstone checkout with [uv](https://docs.astral.sh/uv/getting-started/installation/) installed:
 
 ```sh
-git clone https://github.com/lexmount/browser-eval.git
-cd browser-eval
+git clone https://github.com/lexmount/webstone.git
+cd webstone
 uv sync --frozen
 uv run --frozen python -m harness.benchmarks.data
 uv run --frozen lexbench-webfetch list
 uv run --frozen lexbench-webfetch validate
 ```
 
-Browser Eval downloads the dataset revision selected by its [data lock](https://github.com/lexmount/browser-eval/blob/main/src/harness/benchmarks/data-lock.json) and verifies the archive and file hashes. Public downloads require no access token. The [benchmark integration guide](https://github.com/lexmount/browser-eval/blob/main/docs/benchmarks.md) covers revision selection and offline caching.
+Webstone downloads the dataset revision selected by its [data lock](https://github.com/lexmount/webstone/blob/main/src/harness/benchmarks/data-lock.json) and verifies the archive and file hashes. Public downloads require no access token. The [benchmark integration guide](https://github.com/lexmount/webstone/blob/main/docs/benchmarks.md) covers revision selection and offline caching.
 
 ### Score the 530-task revision from this download
 
@@ -246,15 +246,15 @@ Together they cover all 530 selected tasks. Each directory's `source-inputs.json
 `frozen_task_bindings[<suite>].exported_cases_sha256` binds its resolved cases.
 
 Each file is the frozen dataset bound by the suite's `definitions/evaluation/.../source-inputs.json`.
-Its hashes were verified, and it reproduces every public task row. The Browser Eval
+Its hashes were verified, and it reproduces every public task row. The Webstone
 revision code then applied `primary-content-20261103`, and the result is serialized
 the way `export` writes it. These files are ready to score as published, with the
 revision already applied. Re-running `lexbench-webfetch export --revision
 primary-content-20261103` needs each suite's private frozen parent inputs, including
-source trees that are not in this download. Re-applying the revision also needs Browser Eval at
-[`351cb27b`](https://github.com/lexmount/browser-eval/commit/351cb27b1d9bf82ae81b177874ea5e624ecb00f0)
+source trees that are not in this download. Re-applying the revision also needs Webstone at
+[`351cb27b`](https://github.com/lexmount/webstone/commit/351cb27b1d9bf82ae81b177874ea5e624ecb00f0)
 or later, because its review inherits unchanged entries from `review-550-20261023`.
-Install a suite with Browser Eval's existing importer, then
+Install a suite with Webstone's existing importer, then
 run it with the population:
 
 ```sh
@@ -289,7 +289,7 @@ uv run --frozen lexbench-webfetch export \
   --output /absolute/path/new-input-directory
 ```
 
-The scorer's `--suite webfetch` exports the mixed 150-task suite, including 20 dynamic/challenge examples. For the 430-task edition, select its 130 general-web, document and media IDs plus the three 100-task suites. For the 550-task edition, select all 150 mixed tasks, those three suites and the 100 article tasks (`--suite articles`). Export verifies inputs before writing a new directory. See [reference inputs and access](https://github.com/lexmount/browser-eval/blob/main/docs/reference-inputs.md).
+The scorer's `--suite webfetch` exports the mixed 150-task suite, including 20 dynamic/challenge examples. For the 430-task edition, select its 130 general-web, document and media IDs plus the three 100-task suites. For the 550-task edition, select all 150 mixed tasks, those three suites and the 100 article tasks (`--suite articles`). Export verifies inputs before writing a new directory. See [reference inputs and access](https://github.com/lexmount/webstone/blob/main/docs/reference-inputs.md).
 
 Score a saved response using its complete task definition:
 
@@ -310,39 +310,60 @@ uv run --frozen browser-harness-eval run \
   --case-ids static-simple --output /absolute/path/new-evidence
 ```
 
-The HTTP baseline does not execute JavaScript. For browser setup and a report you can inspect, follow the [runnable local example](https://github.com/lexmount/browser-eval/blob/main/examples/web_fetch/README.md). Formal comparisons use the certified workflow in the [Web Fetch evaluation guide](https://github.com/lexmount/browser-eval/blob/main/src/harness/web_fetch/README.md).
+The HTTP baseline does not execute JavaScript. For browser setup and a report you can inspect, follow the [runnable local example](https://github.com/lexmount/webstone/blob/main/examples/web_fetch/README.md). Formal comparisons use the certified workflow in the [Web Fetch evaluation guide](https://github.com/lexmount/webstone/blob/main/src/harness/web_fetch/README.md).
 
 ## Reporting and limitations
 
 Report passed/evaluated counts for the declared population, unavailable results and the selected benchmark and scorer revisions. Retain failed calls and timeouts. State timing boundaries, retry policy, local or remote execution, and conversion steps; keep all-call timing separate from timing conditional on success. Optional independent verification has its own disclosed scope and does not run automatically.
 
-Live sites change with location, time, session state and access policy. The suite emphasizes task-relevant stable content, but a frozen reference can still become outdated. Investigate source drift before changing a verdict. Updating task requirements or scoring code does not retroactively update an old measurement. Record the immutable dataset commit, task/reference hashes and scorer source identity with each experiment. Detailed collection, correction and reporting rules are maintained in the [Web Fetch guide](https://github.com/lexmount/browser-eval/blob/main/src/harness/web_fetch/README.md).
+Live sites change with location, time, session state and access policy. The suite emphasizes task-relevant stable content, but a frozen reference can still become outdated. Investigate source drift before changing a verdict. Updating task requirements or scoring code does not retroactively update an old measurement. Record the immutable dataset commit, task/reference hashes and scorer source identity with each experiment. Detailed collection, correction and reporting rules are maintained in the [Web Fetch guide](https://github.com/lexmount/webstone/blob/main/src/harness/web_fetch/README.md).
 
 These are curated workloads, not a representative sample of every website. The article suite's initial-load scope does not measure hidden continuations or interaction-dependent content. Text and link checks do not measure visual fidelity or autonomous browser use. This dataset contains task inputs and requirements; it does not ship provider measurements or establish a standing browser ranking.
 
 ## Sources, related benchmarks and rights
 
-LexBench-Web-Fetch was developed as the Web Fetch benchmark component of [Browser Eval](https://github.com/lexmount/browser-eval). Its task definitions retain source URLs and reference bindings. Project-authored material uses Apache-2.0. The reference texts included for evaluation keep their original language and their source terms, and the Apache-2.0 license does not cover them. To request removal or correction, open a discussion in this repository's Community tab. Full captures, screenshots, document and media files, private logs and measured provider outputs are excluded. A public URL or content hash is not permission to redistribute the underlying material. Read [THIRD_PARTY.md](THIRD_PARTY.md), [NOTICE](NOTICE) and the notices included with the selected snapshot.
+LexBench-Web-Fetch was developed as the Web Fetch benchmark component of [Webstone](https://github.com/lexmount/webstone). Its task definitions retain source URLs and reference bindings. Project-authored material uses Apache-2.0. The reference texts included for evaluation keep their original language and their source terms, and the Apache-2.0 license does not cover them. To request removal or correction, open a discussion in this repository's Community tab. Full captures, screenshots, document and media files, private logs and measured provider outputs are excluded. A public URL or content hash is not permission to redistribute the underlying material. Read [THIRD_PARTY.md](THIRD_PARTY.md), [NOTICE](NOTICE) and the notices included with the selected snapshot.
 
 The following projects have separate populations and terms; they are not subsets of this catalog:
 
-- [WebMainBench](https://github.com/opendatalab/WebMainBench), maintained by OpenDataLab, evaluates main-content extraction from saved web pages. Browser Eval's [LexBench-WebMainBench dataset](https://huggingface.co/datasets/Lexmount/LexBench-WebMainBench) records its selected corpus and corrections.
+- [WebMainBench](https://github.com/opendatalab/WebMainBench), maintained by OpenDataLab, evaluates main-content extraction from saved web pages. Webstone's [LexBench-WebMainBench dataset](https://huggingface.co/datasets/Lexmount/LexBench-WebMainBench) records its selected corpus and corrections.
 - [WCXB](https://github.com/Murrough-Foley/web-content-extraction-benchmark) is the Web Content Extraction Benchmark.
 - [WebArena](https://github.com/web-arena-x/webarena) and [WebArena-Verified](https://github.com/ServiceNow/webarena-verified) supply interactive tasks and evaluators used in separate browser workflows.
 - [LexBench-Headless-Browser](https://huggingface.co/datasets/Lexmount/LexBench-Headless-Browser) preserves the task inputs for the [upstream Headless runtime](https://github.com/lexmount/Lexbench-Headless-Browser), which evaluates browser runtime and automation capabilities.
 
-For search evaluation, use [LexBench-Web-Search](https://huggingface.co/datasets/Lexmount/LexBench-Web-Search). [Moli](https://github.com/lexmount/moli) is one browser that can be evaluated with Web Fetch tasks, and [WebFetch CLI](https://github.com/lexmount/webfetch-cli) is a client for a hosted fetch service. Report creation and sharing are covered by [Browser Eval's reporting documentation](https://github.com/lexmount/browser-eval/blob/main/docs/report-hosting.md).
+For search evaluation, use [LexBench-Web-Search](https://huggingface.co/datasets/Lexmount/LexBench-Web-Search). [Moli](https://github.com/lexmount/moli) is one browser that can be evaluated with Web Fetch tasks, and [WebFetch CLI](https://github.com/lexmount/webfetch-cli) is a client for a hosted fetch service. Report creation and sharing are covered by [Webstone's reporting documentation](https://github.com/lexmount/webstone/blob/main/docs/report-hosting.md).
 
 ## Citation
 
-Cite Browser Eval and record the exact dataset revision used in your experiment. The canonical project citation is [CITATION.cff](https://github.com/lexmount/browser-eval/blob/main/CITATION.cff).
+Cite Webstone and record the exact dataset revision used in your experiment. The canonical project citation is [CITATION.cff](https://github.com/lexmount/webstone/blob/main/CITATION.cff).
 
 ```bibtex
-@software{browser_eval_2026,
-  author = {{Browser Eval contributors}},
-  title = {Browser Eval: Benchmarking Browsers and Web Tools for AI Agents},
+@software{webstone_2026,
+  author = {{Webstone contributors}},
+  title = {Webstone: Benchmarking Browsers and Web Tools for AI Agents},
   note = {Evaluation harness},
   year = {2026},
-  url = {https://github.com/lexmount/browser-eval}
+  url = {https://github.com/lexmount/webstone}
 }
 ```
+
+## Main and full content annotations
+
+`content-scopes-530` adds paired `main` and `full` acceptance profiles for the
+existing 530-task `primary-content-20261103` population. It is a separate suite
+selection; the released default population remains 550 tasks with its original
+requirements. Existing tasks, viewer configurations and reference inputs are retained.
+
+The cohort is scored by Webstone [`1287fdf8`](https://github.com/lexmount/webstone/commit/1287fdf859aaac56129933024abdf1246afbcfb7),
+merged from [Webstone PR #268](https://github.com/lexmount/webstone/pull/268).
+Export it with `lexbench-webfetch export --suite content-scopes-530 --source
+<approved-input-directory> --output <new-directory>`. The approved input directory
+must contain the frozen `dataset.json` bound by this suite's `source-inputs.json`.
+Public reference hashes do not replace those inputs.
+
+Use `webstone validate --dataset <exported-dataset.json>
+--content-scope main` or `--content-scope full` before a run. Each run selects one
+scope. The approved input bound by this cohort declares `main` as its default.
+An explicit option overrides a saved scope; annotated tasks without a saved scope
+fall back to `full`. Older suites keep their original rules. The existing manifest tooling pin still describes the
+released edition; the cohort's `implementation` field identifies its scoring code.

THIRD_PARTY.md

@@ -1,6 +1,6 @@
 # Third-party content
 
-The Apache-2.0 license covers project-owned code, documentation and task requirements. Initial definitions and input-checking concepts originate in [Browser Eval](https://github.com/lexmount/browser-eval); its LexMount copyright notice is retained in NOTICE.
+The Apache-2.0 license covers project-owned code, documentation and task requirements. Initial definitions and input-checking concepts originate in [Webstone](https://github.com/lexmount/webstone); its LexMount copyright notice is retained in NOTICE.
 
 ## Reference texts included for evaluation
 

VERSIONING.md

@@ -2,8 +2,8 @@
 
 LexBench-Web-Fetch uses `vMAJOR.MINOR.PATCH` tags on Hugging Face. Each tag names
 one immutable dataset commit. Keep the resolved full commit, task and reference
-hashes, Browser Eval/scorer commit, and execution settings with every experiment.
-Dataset tags are independent of Browser Eval's software version.
+hashes, Webstone/scorer commit, and execution settings with every experiment.
+Dataset tags are independent of Webstone's software version.
 
 ## Release populations
 
@@ -11,6 +11,7 @@ Dataset tags are independent of Browser Eval's software version.
 | --- | ---: | --- | --- |
 | [v1.0.0](https://huggingface.co/datasets/Lexmount/LexBench-Web-Fetch/tree/v1.0.0) | 430 | Earlier population: 100 each for general web, dynamic sites, popular sites and access restrictions, plus 10 embedded documents, 17 media-information tasks and 3 direct-media tasks | The four original suites' task, catalog and source-input hashes match `bench-v1.0.0`; seven viewer groups select the 430-task population |
 | [v2.1.0](https://huggingface.co/datasets/Lexmount/LexBench-Web-Fetch/tree/v2.1.0) | 550 | Expanded catalog: the 430-task population, 100 article tasks and 20 additional dynamic/challenge examples | Uses the existing 2.1.0 catalog declaration and reviewed reference/scoring provenance, following the earlier `bench-v2.0.1` snapshot |
+| [v2.2.0](https://huggingface.co/datasets/Lexmount/LexBench-Web-Fetch/tree/v2.2.0) | 550 | Same population and task hashes as v2.1.0 | Corrects references and acceptance with the reviewed primary-content task scope revision; rescore saved outputs before comparing with v2.1.0 |
 
 The 550-task catalog follows the earlier 430-task population. Adding tasks
 starts a new major series. A later refresh of the 430-task viewer does not make
@@ -65,7 +66,7 @@ The edition's `population_sha256` hashes the sorted list of `suite`, `id` and
 JSON with `ensure_ascii=False`, sorted keys, two-space indentation and a final
 newline before SHA-256 hashing. Viewer JSONL file hashes remain in `files`.
 
-Browser Eval selects datasets through its data lock and import bindings. Update
+Webstone selects datasets through its data lock and import bindings. Update
 those through its existing pinning command using the reviewed full HF commit.
 README links and moving branches are navigation, not execution inputs. Matching
 tags alone do not establish comparability; task/reference hashes, scorer

definitions.tar.gz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:20c943c8bc116c2bae1f3e1ea810ab4a92358e7e632bedb6530c9f7b7480f785
-size 6006987
+oid sha256:62e18ebeec769624fabe696d6d571746a66278657f67de64a03fbecdc6fb2a19
+size 6909656

definitions/evaluation/content-scopes-530/source-inputs.json

@@ -0,0 +1,13 @@
+{
+  "catalog": "dataset-catalog.json",
+  "catalog_sha256": "f3874d8508e4c2d0986c04bc631e17ffd53c1d9385dad1d8b4765e510410644b",
+  "dataset": "dataset.json",
+  "dataset_sha256": "d92958eb95352e72c3774e02df638c52b15b8f325d4713aa1a1eab2072b0e298",
+  "files": {
+    "dataset.json": "d92958eb95352e72c3774e02df638c52b15b8f325d4713aa1a1eab2072b0e298"
+  },
+  "id": "webfetch-content-scopes-530-d92958eb9535",
+  "kind": "frozen_reference_inputs",
+  "schema_version": 2,
+  "trees": {}
+}

definitions/registry.json

@@ -55,6 +55,26 @@
         "dataset-catalog.json": "8e413e683c309674abf3d297df3f378e76d65ead1e2aa343828e4b0e1d90764a",
         "source-inputs.json": "0d9c729f14dae8abb3323a2cf4734736c4733957c9d74d5484033d204642b91b"
       }
+    },
+    {
+      "id": "content-scopes-530",
+      "dataset_id": "webfetch-content-scopes-530",
+      "case_count": 530,
+      "directory": "evaluation/content-scopes-530",
+      "files": {
+        "annotation-provenance.json": "1d414d610700e737f492cf192b2202a1d9f5eeffc17904007e0aac0c644b3a20",
+        "dataset-catalog.json": "f3874d8508e4c2d0986c04bc631e17ffd53c1d9385dad1d8b4765e510410644b",
+        "source-inputs.json": "c76b829b994e8f9cc15cbc1a05018f2cb5135dee6e6f9144e5e3b014434fdbbc",
+        "tasks.json": "6ae76832f1594e410ab36371c7841a5bd6b4fa65dab39bd6496a8f07c5b22b85"
+      },
+      "kind": "annotation_cohort",
+      "source_population": "primary-content-20261103",
+      "implementation": {
+        "pull_request": 268,
+        "repository": "https://github.com/lexmount/webstone",
+        "revision": "1287fdf859aaac56129933024abdf1246afbcfb7",
+        "state": "merged"
+      }
     }
   ]
 }

manifest.json

@@ -1,8 +1,8 @@
 {
   "archive": {
     "path": "definitions.tar.gz",
-    "sha256": "20c943c8bc116c2bae1f3e1ea810ab4a92358e7e632bedb6530c9f7b7480f785",
-    "size": 6006987
+    "sha256": "62e18ebeec769624fabe696d6d571746a66278657f67de64a03fbecdc6fb2a19",
+    "size": 6909656
   },
   "datasets": {
     "access-restrictions": {
@@ -15,6 +15,11 @@
       "path": "evaluation/articles/tasks.json",
       "sha256": "171241ea0bf3cb3f9482de14a7201885733fa717af56d7ea2b28d86a4516a3bb"
     },
+    "content-scopes-530": {
+      "count": 530,
+      "path": "evaluation/content-scopes-530/tasks.json",
+      "sha256": "6ae76832f1594e410ab36371c7841a5bd6b4fa65dab39bd6496a8f07c5b22b85"
+    },
     "dynamic-content": {
       "count": 100,
       "path": "evaluation/dynamic-content-100-moli-webfetch/tasks.json",
@@ -76,9 +81,9 @@
     ],
     "population_sha256": "f1bb4da7d96478dd36ec08d3c340c46dd09b3a2551adab981be020b8fbfaeb96",
     "source_snapshot_revision": "5cc817112ec9777e0a2da43d54859123f85685a3",
-    "tag": "v2.1.0",
+    "tag": "v2.2.0",
     "task_count": 550,
-    "version": "2.1.0"
+    "version": "2.2.0"
   },
   "files": {
     "LICENSE": {
@@ -90,12 +95,12 @@
       "size": 43
     },
     "THIRD_PARTY.md": {
-      "sha256": "29b39660057fd164d2f16810b4daffdf4f0f576dc286634ac837cbd8606ddc8e",
-      "size": 2092
+      "sha256": "128f9a608d95e540b0d2d13b3e4bb07af6b67422902bb82355156c2a6c66c3a5",
+      "size": 2084
     },
     "VERSIONING.md": {
-      "sha256": "fa5f828ac715452cc1a247f39a770916792b38784a835f927a08f00e5b23ab41",
-      "size": 4646
+      "sha256": "6d17c2869b68ce7d433272ffa6cbcfeba5b8f56a1c358c4eef53cc4a74e4b121",
+      "size": 4913
     },
     "data/access-restrictions.jsonl": {
       "sha256": "fc4e0a67760198b3001e85ef8c47062e97243a7072c69a6fd34f96063fd2fca3",
@@ -193,6 +198,22 @@
       "sha256": "171241ea0bf3cb3f9482de14a7201885733fa717af56d7ea2b28d86a4516a3bb",
       "size": 1079224
     },
+    "definitions/evaluation/content-scopes-530/annotation-provenance.json": {
+      "sha256": "1d414d610700e737f492cf192b2202a1d9f5eeffc17904007e0aac0c644b3a20",
+      "size": 527178
+    },
+    "definitions/evaluation/content-scopes-530/dataset-catalog.json": {
+      "sha256": "f3874d8508e4c2d0986c04bc631e17ffd53c1d9385dad1d8b4765e510410644b",
+      "size": 419480
+    },
+    "definitions/evaluation/content-scopes-530/source-inputs.json": {
+      "sha256": "c76b829b994e8f9cc15cbc1a05018f2cb5135dee6e6f9144e5e3b014434fdbbc",
+      "size": 477
+    },
+    "definitions/evaluation/content-scopes-530/tasks.json": {
+      "sha256": "6ae76832f1594e410ab36371c7841a5bd6b4fa65dab39bd6496a8f07c5b22b85",
+      "size": 10045875
+    },
     "definitions/evaluation/dataset-catalog.json": {
       "sha256": "994fa53f7fc7c3170692a7bb493ebc5af9ad8ffb5b576c022fa123e6e6fb3838",
       "size": 100357
@@ -362,13 +383,13 @@
       "size": 50741
     },
     "definitions/registry.json": {
-      "sha256": "f3d1d2a38308b269cd06c4e2f408e8fcd211baae848581ce8532930cf3089a8b",
-      "size": 2411
+      "sha256": "da8a56b46c0419134fc954599bdf4bf6762a7459e1c94e24708b4ef1eaaec323",
+      "size": 3310
     }
   },
   "schema": "browser-eval-dataset-release/1",
   "source_repository": "https://huggingface.co/datasets/Lexmount/LexBench-Web-Fetch",
-  "source_revision": "6b3e1e526600eb91efbb7211c409742b1274888f",
+  "source_revision": "6ae0fc167a4f3c6228703e7ab75c574bd1accb34",
   "track": "webfetch",
   "viewer": {
     "configurations": [

tools/rebuild_publication.py

@@ -68,6 +68,19 @@ def publication_files(root):
     return files
 
 
+
+def dataset_bindings(root):
+    registry = json.loads((root / 'definitions/registry.json').read_text())
+    datasets = {}
+    for row in registry['suites']:
+        path = (Path(row['directory']) / 'tasks.json').as_posix()
+        raw = (root / 'definitions' / path).read_bytes()
+        if sha(raw) != row['files']['tasks.json'] or row['id'] in datasets:
+            raise ValueError('Invalid registry dataset bindings')
+        datasets[row['id']] = {'path': path, 'sha256': sha(raw), 'count': row['case_count']}
+    return datasets
+
+
 def rebuild(root, source_revision):
     if not re.fullmatch(r'[0-9a-f]{40}', source_revision):
         raise ValueError('Use a full uploaded source commit, not a local label')
@@ -80,6 +93,7 @@ def rebuild(root, source_revision):
     manifest['files'] = {name: {'sha256': sha(raw), 'size': len(raw)}
                          for name, raw in sorted(files.items())}
     manifest['source_revision'] = source_revision
+    manifest['datasets'] = dataset_bindings(root)
     # Restore the exact edition object after rebuilding hashes; never invent a tag.
     manifest['edition'] = edition
     buffer = io.BytesIO()
@@ -112,6 +126,8 @@ def check(root):
         raw = path.read_bytes()
         if sha(raw) != binding['sha256'] or len(raw) != binding['size']:
             raise ValueError('Changed publication file: ' + name)
+    if manifest.get('datasets') != dataset_bindings(root):
+        raise ValueError('Manifest dataset bindings differ from the registry')
     raw = (root / 'definitions.tar.gz').read_bytes()
     if sha(raw) != manifest['archive']['sha256'] or len(raw) != manifest['archive']['size']:
         raise ValueError('Changed definitions archive')

tools/test_content_scopes.py

@@ -0,0 +1,26 @@
+"""Annotation cohorts add acceptance scopes without changing the released population."""
+from pathlib import Path
+import json
+import unittest
+from test_review_scope import suite_dirs
+
+ROOT = Path(__file__).resolve().parents[1]
+
+class ContentScopesTests(unittest.TestCase):
+    def test_annotations_reuse_population_and_preserve_both_scopes(self):
+        tasks = json.loads((ROOT/'definitions/evaluation/content-scopes-530/tasks.json').read_text())['cases']
+        population = json.loads((ROOT/'definitions/populations/primary-content-20261103.json').read_text())
+        self.assertEqual(len(tasks), 530)
+        self.assertEqual({c['id'] for c in tasks}, {c['suite']+'--'+c['id'] for c in population['cases']})
+        for task in tasks:
+            self.assertEqual(set(task['content_scopes']) - {'version'}, {'main', 'full'})
+        self.assertNotIn('content-scopes-530', suite_dirs())
+        self.assertEqual(set(suite_dirs()), set(population['frozen_task_bindings']))
+
+    def test_annotation_registry_does_not_expand_edition(self):
+        manifest = json.loads((ROOT/'manifest.json').read_text())
+        self.assertEqual(manifest['edition']['task_count'], 550)
+        registry = json.loads((ROOT/'definitions/registry.json').read_text())
+        entry = next(s for s in registry['suites'] if s['id']=='content-scopes-530')
+        self.assertEqual(entry['kind'], 'annotation_cohort')
+        self.assertEqual(entry['source_population'], 'primary-content-20261103')

tools/test_rebuild_publication.py

@@ -14,6 +14,9 @@ class PublicationTests(unittest.TestCase):
         self.root = Path(self.temp.name)
         (self.root/'definitions').mkdir()
         (self.root/'definitions/tasks.json').write_text('{}\n')
+        registry = {'suites': [{'id': 'test', 'directory': '.', 'case_count': 1,
+                               'files': {'tasks.json': sha(b'{}\n')}}]}
+        (self.root/'definitions/registry.json').write_bytes(encoded(registry))
         for name in ['LICENSE','NOTICE','THIRD_PARTY.md','VERSIONING.md']:
             (self.root/name).write_text(name+'\n')
         (self.root/'data').mkdir()
@@ -83,3 +86,20 @@ class PublicationTests(unittest.TestCase):
 
 if __name__=='__main__':
     unittest.main()
+
+
+class AnnotationManifestTests(unittest.TestCase):
+    setUp = PublicationTests.setUp
+    def test_registered_annotation_is_selectable_without_expanding_edition(self):
+        manifest = rebuild(self.root, 'a' * 40)
+        self.assertEqual(manifest['edition']['task_count'], 1)
+        self.assertEqual(manifest['datasets'], {'test': {
+            'count': 1, 'path': 'tasks.json', 'sha256': sha(b'{}\n')}})
+
+    def test_stale_dataset_selection_is_rejected(self):
+        rebuild(self.root, 'a' * 40)
+        manifest = json.loads((self.root/'manifest.json').read_text())
+        manifest['datasets'] = {}
+        (self.root/'manifest.json').write_bytes(encoded(manifest))
+        with self.assertRaisesRegex(ValueError, 'dataset bindings'):
+            check(self.root)

tools/test_review_scope.py

@@ -18,7 +18,12 @@ QUANTA_NAV = {'Physics', 'Mathematics', 'Biology', 'Computer Science', 'Topics',
 
 def suite_dirs():
     registry = json.loads((ROOT / 'definitions/registry.json').read_text())
-    return {row['id']: ROOT / 'definitions' / row.get('directory', row['id']) for row in registry['suites']}
+    population = json.loads((ROOT / 'definitions/populations/primary-content-20261103.json').read_text())
+    parents = set(population['frozen_task_bindings'])
+    suites = {row['id']: ROOT / 'definitions' / row.get('directory', row['id']) for row in registry['suites']}
+    if not parents.issubset(suites):
+        raise ValueError('Frozen revision parent suite is missing')
+    return {name: suites[name] for name in parents}
 
 
 def cases(path):