diff --git a/README.md b/README.md
index 8d32f6e..639fd04 100644
--- a/README.md
+++ b/README.md
@@ -111,3 +111,10 @@ Other allocation modes remain rejected for those risk extensions.
 Economic comparison requires the explicit [measurement contract](docs/measurement.md).
 Run the [independent financial benchmark suite](docs/financial-benchmarks.md) with pinned
 real repository adapters; synthetic fixture agreement does not certify market validity.
+# Research findings inventory
+
+Run `quant-lab findings --registry path/to/experiments.db` for a compact inventory,
+or add `--json` for recorded scope, every attempt and verified result metrics.
+The command opens the existing registry read-only and preserves failed/retried
+attempts. Missing or altered result files are marked unverified. Execution success
+is not a hypothesis verdict; human research decisions remain in Studio.
diff --git a/src/quant_lab/cli.py b/src/quant_lab/cli.py
index cd6537c..6317616 100644
--- a/src/quant_lab/cli.py
+++ b/src/quant_lab/cli.py
@@ -72,6 +72,16 @@ def cmd_export_html(args: argparse.Namespace) -> int:
     return 0
 
 
+def cmd_findings(args: argparse.Namespace) -> int:
+    from quant_lab.findings import findings_markdown, research_findings
+
+    report = research_findings(Path(args.registry))
+    print(
+        json.dumps(report, indent=2, ensure_ascii=False) if args.json else findings_markdown(report)
+    )
+    return 0
+
+
 def cmd_validate(args: argparse.Namespace) -> int:
     manifest = load_and_validate_standard_run(Path(args.run_dir))
     payload = {
@@ -119,6 +129,11 @@ def build_parser() -> argparse.ArgumentParser:
     p.add_argument("--db", default=str(_default_db()), help="SQLite index path")
     sub = p.add_subparsers(dest="command", required=True)
 
+    findings = sub.add_parser("findings", help="Read all preregistered studies and failed attempts")
+    findings.add_argument("--registry", required=True, help="Existing TrialRegistry SQLite file")
+    findings.add_argument("--json", action="store_true", help="Include scope and verified evidence")
+    findings.set_defaults(func=cmd_findings)
+
     benchmarks = sub.add_parser(
         "financial-benchmarks", help="Run independent real-adapter benchmarks"
     )
diff --git a/src/quant_lab/findings.py b/src/quant_lab/findings.py
new file mode 100644
index 0000000..fbe17cd
--- /dev/null
+++ b/src/quant_lab/findings.py
@@ -0,0 +1,115 @@
+"""Read-only research inventory; execution success is not a research verdict."""
+
+from collections import Counter
+from pathlib import Path
+
+from quant_lab.research import verify_result
+from quant_lab.trials import TrialRegistry, canonical
+
+
+def research_findings(path: Path) -> dict:
+    registry = TrialRegistry(path, read_only=True)
+    studies = []
+    for study_id in registry.study_ids():
+        spec = registry.definition(study_id)
+        definition = spec["definition"]
+        events = registry.history(study_id)
+        starts = {e["attempt_id"]: e for e in events if e["status"] == "running"}
+        latest = {e["attempt_id"]: e for e in events}
+        attempts = []
+        for attempt_id, event in latest.items():
+            row = {**event, "parameters": starts[attempt_id]["payload"]["parameters"]}
+            payload = event["payload"]
+            if event["status"] == "completed" and isinstance(payload.get("result"), str):
+                try:
+                    root = Path(starts[attempt_id]["payload"]["context"]["artifact_root"]).resolve()
+                    target = (root / payload["result"]).resolve()
+                    if root not in target.parents:
+                        raise ValueError("result escapes its registered artifact root")
+                    result = verify_result(target, payload["sha256"])
+                    if result.get("study_id") != study_id or result.get("attempt_id") != attempt_id:
+                        raise ValueError("result identity mismatch")
+                    row["evidence"] = {
+                        "verified": True,
+                        "metrics": result.get("metrics"),
+                        "comparison": result.get("comparison"),
+                        "objective_evaluation": result.get("objective_evaluation"),
+                    }
+                except (KeyError, ValueError, OSError, TypeError) as exc:
+                    row["evidence"] = {"verified": False, "reason": str(exc)}
+            else:
+                row["evidence"] = {"verified": False, "reason": "no file-backed result"}
+            attempts.append(row)
+        attempted = {canonical(s["payload"]["parameters"]) for s in starts.values()}
+        recipe = definition.get("recipe", {})
+        studies.append(
+            {
+                "study_id": study_id,
+                "hypothesis": definition["hypothesis"],
+                "registered_at": spec["registered_at"],
+                "definition_sha256": spec["sha256"],
+                "family_id": definition.get("family_id"),
+                "interval": recipe.get("interval"),
+                "inputs": recipe.get("inputs"),
+                "source": recipe.get("source"),
+                "selection_rule": definition["selection_rule"],
+                "planned_candidates": len(definition["parameters"]),
+                "unattempted_candidates": sum(
+                    canonical(p) not in attempted for p in definition["parameters"]
+                ),
+                "attempt_counts": dict(Counter(e["status"] for e in latest.values())),
+                "attempts": attempts,
+            }
+        )
+    return {
+        "schema": "quant.research-findings/v1",
+        "registry": str(Path(path).resolve()),
+        "policy": "All attempts retained. Execution status does not imply hypothesis support or promotion.",
+        "studies": studies,
+    }
+
+
+def findings_markdown(report: dict) -> str:
+    def cell(value):
+        return (
+            str(value if value is not None else "not recorded")
+            .replace("|", "\\|")
+            .replace("\n", " ")
+        )
+
+    lines = [
+        "# Research findings",
+        "",
+        report["policy"],
+        "",
+        "| Study | Hypothesis | Family | Planned / unattempted | Attempt outcomes |",
+        "| --- | --- | --- | --- | --- |",
+    ]
+    for row in report["studies"]:
+        counts = ", ".join(
+            f"{key}: {value}" for key, value in sorted(row["attempt_counts"].items())
+        )
+        lines.append(
+            "| "
+            + " | ".join(
+                map(
+                    cell,
+                    [
+                        row["study_id"],
+                        row["hypothesis"],
+                        row["family_id"],
+                        f"{row['planned_candidates']} / {row['unattempted_candidates']}",
+                        counts or "not started",
+                    ],
+                )
+            )
+            + " |"
+        )
+    lines += [
+        "",
+        (
+            "DSR/PBO and human decisions are not inferred from run status. "
+            "Use verified family evidence for selection statistics and Studio Research findings for human decisions."
+        ),
+    ]
+    return "\n".join(lines) + "\n"
diff --git a/src/quant_lab/trials.py b/src/quant_lab/trials.py
index 81d43c1..5aa08cb 100644
--- a/src/quant_lab/trials.py
+++ b/src/quant_lab/trials.py
@@ -267,6 +267,11 @@ def history(self, study_id: str) -> list[dict]:
             for a, s, t, p in rows
         ]
 
+    def study_ids(self) -> list[str]:
+        """List registered studies, including those with no attempts."""
+        with self.connect() as db:
+            return [r[0] for r in db.execute("SELECT study_id FROM studies ORDER BY study_id")]
+
     def definition(self, study_id: str) -> dict:
         with self.connect() as db:
             row = db.execute(
diff --git a/tests/test_findings.py b/tests/test_findings.py
new file mode 100644
index 0000000..453a3ee
--- /dev/null
+++ b/tests/test_findings.py
@@ -0,0 +1,58 @@
+import json
+
+import pytest
+
+from quant_lab.cli import main
+from quant_lab.findings import research_findings
+from quant_lab.research import file_hash
+from quant_lab.trials import TrialRegistry
+
+
+def test_findings_include_unstarted_failed_retried_and_tampered_results(tmp_path, capsys):
+    path = tmp_path / "trials.db"
+    registry = TrialRegistry(path)
+    definition = {
+        "hypothesis": "not yet proven",
+        "parameters": [{"w": 1}, {"w": 2}],
+        "code_identity": "abc",
+        "selection_rule": "report all",
+    }
+    registry.register("pending", definition)
+    registry.register("retry", definition)
+    failed = registry.start("retry", {"w": 1})
+    registry.finish(failed, "failed", {"error": "missing data"})
+    attempt = registry.start("retry", {"w": 1}, context={"artifact_root": str(tmp_path)})
+    result = tmp_path / "result.json"
+    result.write_text(
+        json.dumps(
+            {
+                "study_id": "retry",
+                "attempt_id": attempt,
+                "artifacts": {},
+                "metrics": {"sharpe": 0.2},
+            }
+        )
+    )
+    registry.finish(attempt, "completed", {"result": "result.json", "sha256": file_hash(result)})
+    before = path.read_bytes()
+    report = research_findings(path)
+    assert path.read_bytes() == before
+    assert report["studies"][0]["unattempted_candidates"] == 2
+    retry = report["studies"][1]
+    assert retry["attempt_counts"] == {"failed": 1, "completed": 1}
+    assert retry["unattempted_candidates"] == 1
+    assert retry["attempts"][-1]["evidence"]["metrics"] == {"sharpe": 0.2}
+    assert main(["findings", "--registry", str(path), "--json"]) == 0
+    assert json.loads(capsys.readouterr().out) == report
+    assert main(["findings", "--registry", str(path)]) == 0
+    assert "failed: 1" in capsys.readouterr().out
+    result.write_text("{}")
+    evidence = research_findings(path)["studies"][1]["attempts"][-1]["evidence"]
+    assert evidence["verified"] is False and "metrics" not in evidence
+
+
+def test_reader_does_not_create_missing_registry(tmp_path):
+    path = tmp_path / "missing.db"
+    with pytest.raises(FileNotFoundError):
+        research_findings(path)
+    assert not path.exists()