Rebuild BackIntel analyst and benchmark with a Sol analysis specialist

.agents/skills/backintel-benchmark-world/SKILL.md

@@ -0,0 +1,23 @@
+---
+name: backintel-benchmark-world
+description: Reuse BackIntel's local benchmark setup and independent grading conventions.
+---
+
+Status: unreviewed agent-authored knowledge, 2026-11-09.
+
+BackIntel uses LangSmith Evals over a single Deep Agents analyst. See docs/benchmark.md
+for the task specification and README.md for commands. Add storage adapters behind the
+Store protocol; keep native query paths and dialect skills distinct. Local services are
+started one profile at a time through backintel stores to fit this Mac's memory.
+
+Use canonical typed data for independent polars oracles, and store-visible copies for the
+analyst. Keep reference files, loading credentials, grader code and planting declarations
+outside all agent-visible mounts. Preserve source/reference hashes and LangSmith run IDs.
+Reports need explicit metric identifiers so correct numbers cannot pass under swapped
+labels. Replay cutoffs hide future observations, while predictions known at the cutoff
+remain visible. Declared planting and naturally occurring changes are different evidence.
+
+Fix a harness defect before treating its score as agent capability. Keep the original run
+and label the rerun. Read a complete relevant trace, not only aggregate percentages. A
+numerical score does not establish narrative quality; use the independently versioned
+method judge and its written reasons. Cost reports must say whether judge cost is included.

.github/workflows/ci.yml

@@ -0,0 +1,30 @@
+name: ci
+on:
+  push:
+    branches: [main, codex/backintel-v2]
+  pull_request:
+jobs:
+  checks:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+      - run: uv python install 3.12
+      - run: uv sync --frozen
+      - run: uv run ruff check src tests scripts
+      - run: uv run ruff format --check src tests scripts
+      - run: uv run vulture src --min-confidence 100
+      - run: uv run pytest -q
+  secrets:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Scan candidate files with Gitleaks 8.30.1
+        run: |
+          curl -fsSL https://github.com/gitleaks/gitleaks/releases/download/v8.30.1/gitleaks_8.30.1_linux_x64.tar.gz -o /tmp/gitleaks.tar.gz
+          tar -xzf /tmp/gitleaks.tar.gz -C /tmp gitleaks
+          /tmp/gitleaks dir . --redact --no-banner