diff --git a/docs/evidence/ARCHITECTURE_OVERVIEW.md b/docs/evidence/ARCHITECTURE_OVERVIEW.md
new file mode 100644
index 00000000..925daece
--- /dev/null
+++ b/docs/evidence/ARCHITECTURE_OVERVIEW.md
@@ -0,0 +1,29 @@
+# Architecture Overview for Contributors
+
+## High-Level Structure
+```
+src/openamp_foundry/     — pipeline code
+  cli/                   — command-line interface
+  pipeline.py            — main scoring pipeline
+  scoring/               — individual scorers
+  features/              — feature extraction
+  simulation/            — virtual assay modules
+  calibration/           — recalibration engine
+  evidence/              — certificate generation
+tests/                   — test suite
+docs/                    — documentation
+scripts/                 — utility scripts
+configs/                 — configuration files
+schemas/                 — artifact schemas
+```
+
+## Key Files
+- `src/openamp_foundry/pipeline.py` — main scoring pipeline
+- `src/openamp_foundry/cli/main.py` — CLI entry points
+- `configs/pipeline.yaml` — scoring weights and filters
+- `docs/evidence/METRICS_CURRENT.md` — current benchmark metrics
+
+## Data Flow
+1. CLI parses arguments and calls `pipeline.run_ranking_pipeline()`.
+2. Pipeline loads candidates, scores them, and writes outputs.
+3. Outputs include ranked JSONL, evidence certificates, and reports.
diff --git a/docs/evidence/DIAGRAM_ARTIFACT_FLOW.md b/docs/evidence/DIAGRAM_ARTIFACT_FLOW.md
new file mode 100644
index 00000000..ac4b2cbf
--- /dev/null
+++ b/docs/evidence/DIAGRAM_ARTIFACT_FLOW.md
@@ -0,0 +1,19 @@
+# End-to-End Artifact Flow Diagram
+
+```mermaid
+flowchart LR
+  A[Candidate CSV] --> B[score_candidates]
+  B --> C[Ranked JSONL]
+  B --> D[Evidence certs]
+  C --> E[Batch report]
+  C --> F[Run manifest]
+  D --> G[Lab batch pack]
+  G --> H[Lab results]
+  H --> I[calibration-intake]
+  I --> J[recalibration-gate]
+  J -->|pass| K[Weight proposal]
+  J -->|fail| L[Blocked]
+```
+
+This diagram shows the flow from candidate input through scoring, reporting,
+lab partnership, and calibration.
diff --git a/docs/evidence/DIAGRAM_DATA_FLOW.md b/docs/evidence/DIAGRAM_DATA_FLOW.md
new file mode 100644
index 00000000..cf7d7ec7
--- /dev/null
+++ b/docs/evidence/DIAGRAM_DATA_FLOW.md
@@ -0,0 +1,36 @@
+# Data Flow and Data Boundary Map
+
+```mermaid
+flowchart LR
+  subgraph Input
+    A[Candidate CSV]
+    B[Reference CSV]
+    C[Config YAML]
+  end
+  subgraph Pipeline
+    D[score_candidates]
+    E[rank_candidates]
+    F[select_diverse]
+  end
+  subgraph Output
+    G[Ranked JSONL]
+    H[Evidence certs]
+    I[Run manifest]
+    J[Batch report]
+  end
+  A --> D
+  B --> D
+  C --> D
+  D --> E
+  E --> F
+  F --> G
+  D --> H
+  D --> I
+  E --> J
+```
+
+## Data Boundaries
+- **Input boundary**: Candidate CSV, reference CSV, config YAML
+- **Pipeline boundary**: `score_candidates` → `rank_candidates` → `select_diverse`
+- **Output boundary**: JSONL, evidence certs, manifest, report
+- **External boundary**: Lab results (returned by partners)
diff --git a/docs/evidence/DIAGRAM_DEPENDENCY_GRAPH.md b/docs/evidence/DIAGRAM_DEPENDENCY_GRAPH.md
new file mode 100644
index 00000000..f39c5ec7
--- /dev/null
+++ b/docs/evidence/DIAGRAM_DEPENDENCY_GRAPH.md
@@ -0,0 +1,22 @@
+# Dependency Graph Explanation
+
+How pipeline components depend on each other.
+
+```mermaid
+flowchart TD
+  CLI --> Pipeline
+  Pipeline --> Features
+  Pipeline --> Scoring
+  Pipeline --> Evidence
+  Scoring --> Features
+  Scoring --> Ensemble
+  Simulation --> Scoring
+  Calibration --> Pipeline
+```
+
+## Description
+- **CLI** depends on **Pipeline** (calls `run_ranking_pipeline`)
+- **Pipeline** depends on **Features** (feature extraction), **Scoring** (individual scorers), and **Evidence** (certificate generation)
+- **Scoring** depends on **Features** (physicochemical features) and **Ensemble** (weighted combination)
+- **Simulation** feeds into **Scoring** (via `--simulation-mode info`)
+- **Calibration** feeds back into **Pipeline** (weight updates after human review)
diff --git a/docs/evidence/GLOSSARY_BEGINNER_PATH.md b/docs/evidence/GLOSSARY_BEGINNER_PATH.md
new file mode 100644
index 00000000..fd7593b8
--- /dev/null
+++ b/docs/evidence/GLOSSARY_BEGINNER_PATH.md
@@ -0,0 +1,14 @@
+# Beginner Glossary Path
+
+Recommended order to learn project terminology.
+
+1. Candidate — a peptide sequence being evaluated
+2. Score — a numerical prediction of a property
+3. Ensemble — weighted combination of individual scores
+4. Benchmark — a test of pipeline performance against known data
+5. Gate — a safety check that blocks or allows an action
+6. Calibration — adjusting scores based on real data
+7. Evidence certificate — machine-readable record of a candidate's scores
+8. Manifest — record of inputs and outputs for a pipeline run
+9. Ablation — removing a component to measure its contribution
+10. Concept card — short explanation of a single concept
diff --git a/docs/evidence/GUIDE_ARTIFACT_LIFECYCLE.md b/docs/evidence/GUIDE_ARTIFACT_LIFECYCLE.md
new file mode 100644
index 00000000..20366467
--- /dev/null
+++ b/docs/evidence/GUIDE_ARTIFACT_LIFECYCLE.md
@@ -0,0 +1,20 @@
+# Artifact Lifecycle Guide
+
+The lifecycle of a pipeline artifact from creation to archive.
+
+## Stages
+1. **Created** — generated by a pipeline run or script
+2. **Validated** — checked against its schema
+3. **Used** — consumed by downstream processes (reports, calibration)
+4. **Reviewed** — included in a review packet or release
+5. **Archived** — stored for historical reference after its active use
+
+## Rules
+- Every artifact should be validated at creation time.
+- Artifacts should not be modified after creation (immutable).
+- If an artifact needs to change, create a new version.
+- Artifacts used in releases should be archived with the release.
+
+## Related
+- `docs/evidence/ARTIFACT_MAP.md` — artifact catalog
+- `docs/evidence/STANDARD_GENERATED_FILE_WARNING.md` — generated file warnings
diff --git a/docs/evidence/GUIDE_MAKE_TARGETS.md b/docs/evidence/GUIDE_MAKE_TARGETS.md
new file mode 100644
index 00000000..8583cd92
--- /dev/null
+++ b/docs/evidence/GUIDE_MAKE_TARGETS.md
@@ -0,0 +1,20 @@
+# Make Target Guide
+
+How to use Makefile targets in the project.
+
+## Common Targets
+| Target | Purpose |
+|--------|---------|
+| `make demo` | Run the demo pipeline |
+| `make test` | Run all tests |
+| `make bench-500` | Run 500-AMP benchmark |
+| `make bench-gate` | Check benchmark regression gates |
+| `make regenerate-all` | Regenerate all outputs |
+| `make full-reproducibility-report` | Generate release report |
+| `make lab-batch-pack` | Build lab partner zip |
+
+## Adding a New Target
+1. Add the target to the Makefile.
+2. Add it to the `.PHONY` list if it doesn't create a file.
+3. Add a description to the help section.
+4. Document the target in COMMAND_SURFACE.md if it's user-facing.
diff --git a/docs/evidence/GUIDE_QUICKSTART_DIAGNOSIS.md b/docs/evidence/GUIDE_QUICKSTART_DIAGNOSIS.md
new file mode 100644
index 00000000..498ab872
--- /dev/null
+++ b/docs/evidence/GUIDE_QUICKSTART_DIAGNOSIS.md
@@ -0,0 +1,12 @@
+# Failed Quickstart Diagnosis Table
+
+Common quickstart failures and how to fix them.
+
+| Symptom | Likely Cause | Solution |
+|---------|-------------|----------|
+| `command not found: openamp-foundry` | Package not installed | `pip install -e ".[dev]"` |
+| `ModuleNotFoundError: No module named 'openamp_foundry'` | PYTHONPATH not set | `export PYTHONPATH=src` |
+| `FileNotFoundError: examples/...` | Wrong working directory | Run from repo root |
+| Tests fail | Wrong Python version | `python3 --version` (requires 3.11) |
+| `make demo` produces no output | Previous run cleaned up | Re-run `make demo` |
+| Git error | Not in a git repo | `git clone` first |
diff --git a/docs/evidence/GUIDE_QUICKSTART_OUTPUT_GALLERY.md b/docs/evidence/GUIDE_QUICKSTART_OUTPUT_GALLERY.md
new file mode 100644
index 00000000..b39219f1
--- /dev/null
+++ b/docs/evidence/GUIDE_QUICKSTART_OUTPUT_GALLERY.md
@@ -0,0 +1,21 @@
+# Expected Output Gallery for Quickstart
+
+After running `make demo`, you should see these outputs.
+
+## Files Created
+```
+outputs/demo_ranked.jsonl    — ranked candidate list
+outputs/demo_report.md       — human-readable report
+outputs/evidence/            — evidence certificates directory
+outputs/evidence/AMPF-*.json — one certificate per candidate
+outputs/run_manifest.json    — run manifest
+```
+
+## Sample Output
+The first line of `outputs/demo_ranked.jsonl` should contain a JSON object
+with `candidate_id`, `sequence`, `scores`, and `features` fields.
+
+## If Output Differs
+- Check that `make demo` completed without errors.
+- Check that the input files exist at `examples/sequences/demo_candidates.csv`.
+- Check that the reference file exists at `examples/known_reference/demo_known_amps.csv`.
diff --git a/docs/evidence/MAP_COMMAND_TO_ARTIFACT.md b/docs/evidence/MAP_COMMAND_TO_ARTIFACT.md
new file mode 100644
index 00000000..99516c47
--- /dev/null
+++ b/docs/evidence/MAP_COMMAND_TO_ARTIFACT.md
@@ -0,0 +1,15 @@
+# Command-to-Artifact Map
+
+Maps CLI commands to their output artifacts.
+
+| Command | Output Artifact | Format | Location |
+|---------|----------------|:-----:|----------|
+| `rank` | Ranked candidates | JSONL | `outputs/*_ranked.jsonl` |
+| `rank` | Evidence certificates | JSON | `outputs/evidence/` |
+| `rank` | Run manifest | JSON | `outputs/run_manifest.json` |
+| `rank --report` | Batch report | MD + JSON | `outputs/*_report.md` |
+| `validate` | Validation result | stdout | — |
+| `pilot-panel` | Pilot panel CSV | CSV | `outputs/*_panel.csv` |
+| `lab-result-report` | Lab result report | JSON | `outputs/*_report.json` |
+| `calibration-intake` | Intake report | JSON | `outputs/*_intake.json` |
+| `recalibration-gate` | Gate verdict | JSON | `outputs/*_verdict.json` |
diff --git a/docs/evidence/POLICY_CLI_REFERENCE_GENERATION.md b/docs/evidence/POLICY_CLI_REFERENCE_GENERATION.md
new file mode 100644
index 00000000..b6a2ee51
--- /dev/null
+++ b/docs/evidence/POLICY_CLI_REFERENCE_GENERATION.md
@@ -0,0 +1,14 @@
+# CLI Reference Generation Policy
+
+CLI reference docs should be generated from the code, not written manually.
+
+## Current State
+CLI reference is maintained in docs/getting-started/COMMAND_SURFACE.md.
+
+## Policy
+- Add new commands to COMMAND_SURFACE.md when they are added.
+- Update COMMAND_SURFACE.md when command flags change.
+- Consider auto-generating the reference from argparse in the future.
+
+## Related
+- `docs/getting-started/COMMAND_SURFACE.md` — CLI reference
diff --git a/tests/test_arch_quickstart_batch.py b/tests/test_arch_quickstart_batch.py
new file mode 100644
index 00000000..efee261e
--- /dev/null
+++ b/tests/test_arch_quickstart_batch.py
@@ -0,0 +1,21 @@
+"""Tests for architecture and quickstart batch."""
+from pathlib import Path
+
+FILES = [
+    "POLICY_CLI_REFERENCE_GENERATION.md",
+    "GUIDE_MAKE_TARGETS.md",
+    "GUIDE_QUICKSTART_OUTPUT_GALLERY.md",
+    "GUIDE_QUICKSTART_DIAGNOSIS.md",
+    "GLOSSARY_BEGINNER_PATH.md",
+    "DIAGRAM_ARTIFACT_FLOW.md",
+    "ARCHITECTURE_OVERVIEW.md",
+    "GUIDE_ARTIFACT_LIFECYCLE.md",
+    "DIAGRAM_DATA_FLOW.md",
+    "MAP_COMMAND_TO_ARTIFACT.md",
+    "DIAGRAM_DEPENDENCY_GRAPH.md",
+]
+
+
+def test_all_exist():
+    for f in FILES:
+        assert Path("docs/evidence", f).exists(), f"Missing: {f}"