Add WAMProbe counterfactual MVP and evidence map

.gitattributes

@@ -0,0 +1 @@
+environments/starwam/patches/*.patch -whitespace

.github/ISSUE_TEMPLATE/adapter_proposal.yml

@@ -0,0 +1,42 @@
+name: Adapter proposal
+description: Propose integration of an open World Action Model
+title: "[Adapter]: "
+labels: ["area:adapter"]
+body:
+  - type: input
+    id: repository
+    attributes:
+      label: Upstream repository
+      placeholder: https://github.com/owner/repository
+    validations:
+      required: true
+  - type: input
+    id: checkpoint
+    attributes:
+      label: Public checkpoint
+    validations:
+      required: true
+  - type: input
+    id: license
+    attributes:
+      label: Code and checkpoint licenses
+    validations:
+      required: true
+  - type: checkboxes
+    id: capabilities
+    attributes:
+      label: Exposed capabilities
+      options:
+        - label: Pixel future prediction
+        - label: State future prediction
+        - label: Latent future representation
+        - label: Action prediction
+        - label: Candidate action scoring
+        - label: World feature access
+  - type: textarea
+    id: smoke
+    attributes:
+      label: Smallest reproducible smoke path
+      description: Describe hardware, benchmark, and expected runtime.
+    validations:
+      required: true

.github/ISSUE_TEMPLATE/benchmark_proposal.yml

@@ -0,0 +1,51 @@
+name: Benchmark proposal
+description: Propose a paired intervention benchmark or benchmark extension
+title: "[Benchmark]: "
+labels: ["area:benchmark", "needs-rfc"]
+body:
+  - type: textarea
+    id: diagnostic_scope
+    attributes:
+      label: Diagnostic scope
+      description: Name the WAM failure modes and capability tier this benchmark evaluates.
+    validations:
+      required: true
+  - type: textarea
+    id: intervention_contract
+    attributes:
+      label: Intervention and restore contract
+      description: Describe shared contexts, legal action branches, snapshot restore, and branch-order checks.
+    validations:
+      required: true
+  - type: textarea
+    id: data_license
+    attributes:
+      label: Data, simulator, and asset licenses
+      description: Include upstream versions, licenses, redistribution limits, and required hashes.
+    validations:
+      required: true
+  - type: textarea
+    id: splits
+    attributes:
+      label: Splits and leakage controls
+      description: Define ID/OOD splits and how train/evaluation overlap is prevented or reported.
+    validations:
+      required: true
+  - type: textarea
+    id: smoke
+    attributes:
+      label: Smallest reproducible smoke path
+      description: Give one bounded command, hardware requirements, runtime, and expected structured output.
+    validations:
+      required: true
+  - type: checkboxes
+    id: evidence
+    attributes:
+      label: Required evidence
+      options:
+        - label: Generation or conversion code is included.
+          required: true
+        - label: A benchmark card states supported and unsupported claims.
+          required: true
+        - label: Oracle and failure baselines have expected-ordering tests.
+          required: true

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,35 @@
+name: Bug report
+description: Report reproducible incorrect behavior in WAMProbe
+title: "[Bug]: "
+labels: ["bug"]
+body:
+  - type: markdown
+    attributes:
+      value: "Do not include credentials, private datasets, or model access tokens."
+  - type: textarea
+    id: description
+    attributes:
+      label: What happened?
+      description: Describe the observed and expected behavior.
+    validations:
+      required: true
+  - type: textarea
+    id: reproduce
+    attributes:
+      label: Minimal reproduction
+      description: Include the command, resolved config, and smallest non-sensitive fixture.
+    validations:
+      required: true
+  - type: input
+    id: version
+    attributes:
+      label: WAMProbe version or commit
+    validations:
+      required: true
+  - type: textarea
+    id: environment
+    attributes:
+      label: Environment
+      description: Python, OS, hardware, and relevant adapter/upstream revisions.
+    validations:
+      required: true

.github/ISSUE_TEMPLATE/config.yml

@@ -0,0 +1,5 @@
+blank_issues_enabled: true
+contact_links:
+  - name: Security report
+    url: https://github.com/myheart521/WAMProbe/security/advisories/new
+    about: Report vulnerabilities or sensitive security concerns privately.

.github/ISSUE_TEMPLATE/metric_proposal.yml

@@ -0,0 +1,32 @@
+name: Metric proposal
+description: Propose a causal, dynamics, utility, uncertainty, or efficiency metric
+title: "[Metric]: "
+labels: ["area:metric", "needs-rfc"]
+body:
+  - type: textarea
+    id: failure
+    attributes:
+      label: Failure mode
+      description: What concrete WAM failure does this metric detect?
+    validations:
+      required: true
+  - type: textarea
+    id: definition
+    attributes:
+      label: Definition and capability requirements
+      description: Include directionality, aggregation unit, and stochastic-model handling.
+    validations:
+      required: true
+  - type: textarea
+    id: gaming
+    attributes:
+      label: How can the metric be gamed?
+    validations:
+      required: true
+  - type: textarea
+    id: baselines
+    attributes:
+      label: Expected reference-baseline behavior
+      description: State the expected ordering for oracle, copy-last, wrong-direction, and action-agnostic baselines.
+    validations:
+      required: true

.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,22 @@
+## What changed
+
+<!-- Summarize the implementation and public behavior. -->
+
+## Why
+
+<!-- Explain the scientific or engineering failure mode addressed. -->
+
+## Validation
+
+- [ ] `ruff format --check .`
+- [ ] `ruff check .`
+- [ ] `mypy`
+- [ ] `python scripts/validate_repository.py`
+- [ ] `mkdocs build --strict`
+- [ ] `pytest --cov=wamprobe --cov-report=term-missing`
+
+## Metric/adapter/benchmark checklist
+
+- [ ] Not applicable, or capability requirements are documented.
+- [ ] Not applicable, or expected reference-baseline behavior is tested.
+- [ ] Not applicable, or data/checkpoint license and upstream revision are documented.

.github/dependabot.yml

@@ -0,0 +1,12 @@
+version: 2
+updates:
+  - package-ecosystem: pip
+    directory: /
+    schedule:
+      interval: monthly
+    open-pull-requests-limit: 5
+  - package-ecosystem: github-actions
+    directory: /
+    schedule:
+      interval: monthly
+    open-pull-requests-limit: 5

.github/workflows/ci.yml

@@ -0,0 +1,59 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+permissions:
+  contents: read
+
+jobs:
+  quality:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.11", "3.12", "3.13"]
+
+    steps:
+      - uses: actions/checkout@v7
+      - uses: actions/setup-python@v6
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: pip
+      - name: Install project
+        run: python -m pip install -e '.[dev]'
+      - name: Check formatting
+        run: ruff format --check .
+      - name: Lint
+        run: ruff check .
+      - name: Type check
+        run: mypy
+      - name: Validate JSON Schemas and Markdown links
+        if: matrix.python-version == '3.11'
+        run: python scripts/validate_repository.py
+      - name: Build documentation site
+        if: matrix.python-version == '3.11'
+        run: mkdocs build --strict
+      - name: Test with coverage
+        run: pytest --cov=wamprobe --cov-report=term-missing --cov-fail-under=85
+
+  package-candidate:
+    needs: quality
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v7
+        with:
+          fetch-depth: 0
+      - uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+      - uses: astral-sh/setup-uv@v8.3.2
+        with:
+          version: "0.11.6"
+          enable-cache: true
+      - name: Install locked development environment
+        run: uv sync --extra dev --locked
+      - name: Build, reproduce, audit, and smoke candidate
+        run: .venv/bin/python scripts/build_release_candidate.py --dist-dir dist/ci-candidate

.github/workflows/docs.yml

@@ -0,0 +1,47 @@
+name: Documentation
+
+on:
+  push:
+    branches: [main]
+    paths:
+      - "docs/**"
+      - "mkdocs.yml"
+      - "pyproject.toml"
+      - "uv.lock"
+      - ".github/workflows/docs.yml"
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  pages: write
+  id-token: write
+
+concurrency:
+  group: github-pages
+  cancel-in-progress: true
+
+jobs:
+  build-and-deploy:
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v7
+      - uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+          cache: pip
+      - name: Install documentation environment
+        run: python -m pip install -e '.[dev]'
+      - name: Validate repository documentation
+        run: python scripts/validate_repository.py
+      - name: Build documentation site
+        run: mkdocs build --strict
+      - uses: actions/configure-pages@v6
+      - uses: actions/upload-pages-artifact@v5
+        with:
+          path: site
+      - name: Deploy to GitHub Pages
+        id: deployment
+        uses: actions/deploy-pages@v5

.github/workflows/gpu-nightly.yml

@@ -0,0 +1,80 @@
+name: GPU nightly
+
+on:
+  schedule:
+    - cron: "17 3 * * *"
+  workflow_dispatch:
+    inputs:
+      full_matrix:
+        description: Run all four tasks, seeds, and NFE settings
+        required: false
+        default: false
+        type: boolean
+
+permissions:
+  contents: read
+
+concurrency:
+  group: wamprobe-gpu-nightly
+  cancel-in-progress: false
+
+jobs:
+  starwam-libero-smoke:
+    if: ${{ vars.WAMPROBE_GPU_NIGHTLY_ENABLED == 'true' }}
+    runs-on: [self-hosted, linux, x64, gpu]
+    timeout-minutes: 90
+
+    steps:
+      - uses: actions/checkout@v7
+        with:
+          clean: false
+      - name: Validate provisioned runner
+        run: |
+          test -x environments/starwam/.venv/bin/python
+          test -d vendor/upstream/starwam
+          test -d vendor/upstream/libero
+          nvidia-smi
+          environments/starwam/.venv/bin/python -m wamprobe doctor --json
+      - name: Run cached StarWAM matrix
+        env:
+          FULL_MATRIX: ${{ inputs.full_matrix || 'false' }}
+        run: |
+          args=(
+            --mode matrix
+            --gpu-index 0
+            --minimum-free-gib 18
+            --vae-device cpu
+            --run-dir runs/gpu-nightly
+          )
+          if [[ "$FULL_MATRIX" != "true" ]]; then
+            args+=(
+              --task-key spatial-task0
+              --matrix-seed 0
+              --matrix-inference-steps 1
+            )
+          fi
+          environments/starwam/.venv/bin/python \
+            environments/starwam/run_libero_smoke.py "${args[@]}"
+      - name: Execute action prefixes
+        run: |
+          environments/starwam/.venv/bin/python \
+            environments/starwam/execute_prediction_matrix.py \
+            --gpu-index 0 \
+            --matrix-dir runs/gpu-nightly
+      - name: Build experiment report
+        run: |
+          environments/starwam/.venv/bin/python -m wamprobe experiment-report \
+            runs/gpu-nightly/matrix-index.json \
+            runs/gpu-nightly/execution-index.json \
+            --output runs/gpu-nightly/report
+      - name: Upload structured evidence
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v7
+        with:
+          name: wamprobe-gpu-nightly-${{ github.run_id }}
+          if-no-files-found: warn
+          retention-days: 14
+          path: |
+            runs/gpu-nightly/matrix-index.json
+            runs/gpu-nightly/execution-index.json
+            runs/gpu-nightly/report/