I'll work through the operating mode detection and setup before any analysis. --- ``` OPERATING MODE: 1 — Tree-First Justification: INPUT_AVAILABLE = tree. No pipeline files, dbt model SQL, Great Expectations suites, or test configs are in context. All coverage findings below are {Speculative: tree-only} until Phase 2 files are supplied. QUALITY DIMENSIONS IN SCOPE: Completeness | Validity | Uniqueness (Accuracy and Timeliness excluded per QUALITY_PRIORITY; flagged if structural signals make gaps visible) FILES ASSESSED (Direct): NONE FILES IN TREE ONLY (Speculative): awaiting your tree output FILES NOT AVAILABLE: all pipeline files, dbt models, GE suites, sources.yml, schema.yml test blocks, Airflow DAG definitions ``` --- ## ⚠️ Tree Not Yet Received You've declared `[INPUT_AVAILABLE] = tree` but no tree output is present in your message. I cannot speculate on folder topology, naming conventions, or coverage gaps without it. **Do not proceed to analysis until the tree is provided.** Fabricating structure would violate the strict constraints of this framework. --- ## What to Paste Run one of the following in your project root and paste the full output: **Option A — Standard Unix tree** ```bash tree -a -I '__pycache__|*.pyc|.git|.venv|node_modules' \ --dirsfirst -L 5 ``` **Option B — find fallback (if tree is not installed)** ```bash find . -type f \ \( -name "*.py" -o -name "*.sql" -o -name "*.yml" \ -o -name "*.yaml" -o -name "*.json" \) \ | grep -v -E '(__pycache__|\.git|\.venv)' \ | sort ``` --- ## What I Will Do With It (Phase 1 Preview) Once the tree arrives, I will: 1. **Map your staging + marts layer topology** — identify extract, transform, and load steps by folder and file naming conventions 2. **Locate test coverage signals** — `schema.yml`, `tests/` directories, GE checkpoint YAMLs, `expectations/` folders, `sources.yml` freshness blocks 3. **Flag unguarded high-risk nodes** — models with likely Fan-Out > 3 and no visible test file sibling 4. **Identify the 5–10 highest-priority files** to request for Phase 2 Direct analysis, ranked by: - Steps feeding downstream mart models (high blast radius) - Staging entry points with no adjacent `schema.yml` - Any load-to-Snowflake step with no visible assertion layer 5. **Produce a speculative coverage map** clearly marked `{Speculative: tree-only}` with `{Confidence: LOW | MEDIUM}` on every inference --- Paste your tree output and Phase 1 analysis begins immediately.