diff --git a/.github/workflows/eval.yml b/.github/workflows/eval.yml
index f92de2a61..c8bbf6e34 100644
--- a/.github/workflows/eval.yml
+++ b/.github/workflows/eval.yml
@@ -50,15 +50,7 @@ jobs:
 
       - name: Download datasets
         run: |
-          lenskit data fetch -D data --movielens ml-100k ml-1m ml-10m
-          lenskit data fetch -D data/az23 --amazon --core Video_Games
-          lenskit data fetch -D data/az18 --amazon --edition 2018 Video_Games
-          lenskit data fetch -D data/az14 --amazon --edition 2014 Video_Games
-          lenskit data fetch --ms-web
-          if [[ ! -f data/australian_users_items.json.gz ]]; then
-            curl -fsL -o data/australian_users_items.json.gz.tmp https://mcauleylab.ucsd.edu/public_datasets/data/steam/australian_users_items.json.gz
-            mv data/australian_users_items.json.gz.tmp data/australian_users_items.json.gz
-          fi
+          lenskit data fetch -D data --movielens ml-100k
 
       - name: Run Eval Tests
         run: |
@@ -103,15 +95,7 @@ jobs:
 
       - name: Download datasets
         run: |
-          coverage run -a --source=src/lenskit -m lenskit data fetch -D data --movielens ml-100k ml-20m ml-1m ml-10m
-          coverage run -a --source=src/lenskit -m lenskit data fetch -D data/az23 --amazon --core Video_Games
-          coverage run -a --source=src/lenskit -m lenskit data fetch -D data/az18 --amazon --edition 2018 Video_Games
-          coverage run -a --source=src/lenskit -m lenskit data fetch -D data/az14 --amazon --edition 2014 Video_Games
-          coverage run -a --source=src/lenskit -m lenskit data fetch --ms-web
-          if [[ ! -f data/australian_users_items.json.gz ]]; then
-            curl -fsL -o data/australian_users_items.json.gz.tmp https://mcauleylab.ucsd.edu/public_datasets/data/steam/australian_users_items.json.gz
-            mv data/australian_users_items.json.gz.tmp data/australian_users_items.json.gz
-          fi
+          coverage run -a --source=src/lenskit -m lenskit data fetch -D data --movielens ml-100k
 
       - name: 🕺🏻 Test LensKit CLI
         run: |
diff --git a/.woodpecker/dataset-tests.yml b/.woodpecker/dataset-tests.yml
new file mode 100644
index 000000000..2969de2c0
--- /dev/null
+++ b/.woodpecker/dataset-tests.yml
@@ -0,0 +1,23 @@
+when:
+  - event: push
+    branch: main
+  - event: pull_request
+
+labels:
+  backend: docker
+
+steps:
+  - name: build and run data tests
+    image: codeberg.org/mdekstrand/woodpecker-mise:v1-f44-dev
+    pull: true
+    settings:
+      task: "ci:prepare ::: test -v --coverage -m realdata ::: test-cli tests/cli/test-data-convert.sh"
+      mise-env: ci
+
+  - name: upload code coverage
+    image: woodpeckerci/plugin-codecov
+    settings:
+      files:
+        - coverage.xml
+      token:
+        from_secret: codecov-token
diff --git a/LICENSE.md b/LICENSE.md
index bc0f03d53..2342ec1e4 100644
--- a/LICENSE.md
+++ b/LICENSE.md
@@ -1,3 +1,8 @@
+> [!NOTE]
+> The following license applies to the LensKit source code. The data sets
+> under `data/` have their own terms, described in their respective README
+> files.
+
 Copyright (c) 2018–2023 Boise State University
 Copyright (c) 2023-2026 Drexel University and contributors
 
diff --git a/codecov.yml b/codecov.yml
index c18d3d1fd..14a84db35 100644
--- a/codecov.yml
+++ b/codecov.yml
@@ -2,6 +2,7 @@ codecov:
   notify:
     after_n_builds: 10
     wait_for_ci: true
+  token: ab58c9cf-25b8-4283-a485-0b6382dc9a61
 ignore:
   - build*.py
   - build-tools/*
diff --git a/mise/config.ci.toml b/mise/config.ci.toml
new file mode 100644
index 000000000..ec0d92c3e
--- /dev/null
+++ b/mise/config.ci.toml
@@ -0,0 +1,5 @@
+[settings]
+disable_tools = ["pkl", "perl"]
+
+[settings.python]
+uv_venv_auto = "source"
diff --git a/mise/config.toml b/mise/config.toml
index fd420ee46..5c447e532 100644
--- a/mise/config.toml
+++ b/mise/config.toml
@@ -2,6 +2,7 @@
 rust = "stable"
 usage = "latest"
 uv = "latest"
+jq = "latest"
 shfmt = "latest"
 tombi = "latest"
 perl = { version = "latest", os = ["linux", "macos"] }
@@ -12,6 +13,7 @@ yamlfmt = "latest"
 ruff = "latest"
 
 [settings]
+experimental = true
 idiomatic_version_file_enable_tools = ["rust"]
 github.github_attestations = false
 aqua.github_attestations = false
@@ -19,8 +21,14 @@ aqua.github_attestations = false
 [settings.python]
 # uv_venv_auto = "source"
 
+[deps.uv]
+
 [env]
 HK_MISE = 1
 
 [hooks]
-postinstall = "hk install"
+postinstall = '''
+if [ -z "$CI" ]; then
+  hk install
+fi
+'''
diff --git a/mise/mise.lock b/mise/mise.lock
index 556f0d2cc..8319455d4 100644
--- a/mise/mise.lock
+++ b/mise/mise.lock
@@ -34,9 +34,44 @@ checksum = "sha256:a2e5c6131d01cfd6d62204872798f97fc933e7f142d718b3e191434e2d31e
 url = "https://github.com/jdx/hk/releases/download/v1.53.0/hk-x86_64-pc-windows-msvc.zip"
 url_api = "https://api.github.com/repos/jdx/hk/releases/assets/487429261"
 
-[[tools.oxfmt]]
-version = "0.60.0"
-backend = "npm:oxfmt"
+[[tools.jq]]
+version = "1.8.2"
+backend = "aqua:jqlang/jq"
+
+[tools.jq."platforms.linux-arm64"]
+checksum = "sha256:8b85c817833814ddca00a144c33705546355afccf0cf39b188f3cdb48b852309"
+url = "https://github.com/jqlang/jq/releases/download/jq-1.8.2/jq-linux-arm64"
+url_api = "https://api.github.com/repos/jqlang/jq/releases/assets/453012756"
+
+[tools.jq."platforms.linux-arm64-musl"]
+checksum = "sha256:8b85c817833814ddca00a144c33705546355afccf0cf39b188f3cdb48b852309"
+url = "https://github.com/jqlang/jq/releases/download/jq-1.8.2/jq-linux-arm64"
+url_api = "https://api.github.com/repos/jqlang/jq/releases/assets/453012756"
+
+[tools.jq."platforms.linux-x64"]
+checksum = "sha256:b1c22172dd303f3be49e935aa56aa48a8b7a46e0bc838b4997d3bb451495870f"
+url = "https://github.com/jqlang/jq/releases/download/jq-1.8.2/jq-linux-amd64"
+url_api = "https://api.github.com/repos/jqlang/jq/releases/assets/453012752"
+
+[tools.jq."platforms.linux-x64-musl"]
+checksum = "sha256:b1c22172dd303f3be49e935aa56aa48a8b7a46e0bc838b4997d3bb451495870f"
+url = "https://github.com/jqlang/jq/releases/download/jq-1.8.2/jq-linux-amd64"
+url_api = "https://api.github.com/repos/jqlang/jq/releases/assets/453012752"
+
+[tools.jq."platforms.macos-arm64"]
+checksum = "sha256:2d75340ba57a4b4b4c8708a21c2dc8e958a48aaa8bba13b27f77f6e4c0eca07e"
+url = "https://github.com/jqlang/jq/releases/download/jq-1.8.2/jq-macos-arm64"
+url_api = "https://api.github.com/repos/jqlang/jq/releases/assets/453012783"
+
+[tools.jq."platforms.macos-x64"]
+checksum = "sha256:e94b266e3c26690550006abe63152b782280f4e14374accdf04cbde844f00bc0"
+url = "https://github.com/jqlang/jq/releases/download/jq-1.8.2/jq-macos-amd64"
+url_api = "https://api.github.com/repos/jqlang/jq/releases/assets/453012782"
+
+[tools.jq."platforms.windows-x64"]
+checksum = "sha256:a6fc67fedaf9128a3309a1e2ebb8b986aeccf70122ee46d2cb4849e423f0c627"
+url = "https://github.com/jqlang/jq/releases/download/jq-1.8.2/jq-windows-amd64.exe"
+url_api = "https://api.github.com/repos/jqlang/jq/releases/assets/453012788"
 
 [[tools.perl]]
 version = "5.44.0.0"
diff --git a/mise/tasks/ci/prepare.sh b/mise/tasks/ci/prepare.sh
new file mode 100755
index 000000000..523aae702
--- /dev/null
+++ b/mise/tasks/ci/prepare.sh
@@ -0,0 +1,23 @@
+#!/bin/zsh
+#MISE description="Prepare CI environment"
+#USAGE flag "-v --verbose" help="Enable verbose logging."
+
+. "$MISE_PROJECT_ROOT/mise/task-functions.sh"
+. "${UV_PROJECT_ENVIRONMENT:-$MISE_PROJECT_ROOT/.venv}/bin/activate"
+
+if [[ -z $CI ]]; then
+    msg "not in CI, skipping environment setup"
+fi
+
+if [[ $CI_SYSTEM_NAME = woodpecker ]]; then
+    if [[ -d /datasets ]]; then
+        msg "linking data from /datasets"
+        for f in /datasets/*; do
+            ln -s $f data/${f#/datasets/}
+        done
+    else
+        msg -warn "running on Woodpecker but /datasets does not exist"
+    fi
+else
+    msg "nothing to do for CI $CI_SYSTEM_NAME"
+fi
diff --git a/mise/tasks/test-cli.sh b/mise/tasks/test-cli.sh
index 7a144d1ba..3422db94f 100755
--- a/mise/tasks/test-cli.sh
+++ b/mise/tasks/test-cli.sh
@@ -1,4 +1,5 @@
 #!/usr/bin/env bash
 #MISE description="Run CLI tests"
+#MISE wait_for=["test"]
 
 exec ./tests/cli/run.sh "$@"
diff --git a/mise/tasks/test.sh b/mise/tasks/test.sh
index add734d5f..eba8dac69 100755
--- a/mise/tasks/test.sh
+++ b/mise/tasks/test.sh
@@ -1,5 +1,6 @@
 #!/usr/bin/env bash
 #MISE description="Run tests"
+#MISE wait_for=["ci:prepare"]
 #USAGE flag "--coverage" help="Run tests with coverage."
 #USAGE flag "--accel-coverage" help="Run tests with Rust accelerator coverage."
 #USAGE flag "--slow" help="Include slow tests."