diff --git a/wizard-work/sol/triage-canonical.json b/wizard-work/sol/triage-canonical.json
new file mode 100644
index 0000000..df4a270
--- /dev/null
+++ b/wizard-work/sol/triage-canonical.json
@@ -0,0 +1,42 @@
+[
+  {
+    "dir": "beam",
+    "state": "done",
+    "file_count": 22
+  },
+  {
+    "dir": "forge",
+    "state": "in-progress",
+    "file_count": 20
+  },
+  {
+    "dir": "ledger",
+    "state": "completed",
+    "file_count": 40
+  },
+  {
+    "dir": "null",
+    "state": "framework",
+    "file_count": 34
+  },
+  {
+    "dir": "quill",
+    "state": "in-progress",
+    "file_count": 1
+  },
+  {
+    "dir": "sol",
+    "state": "done",
+    "file_count": 30
+  },
+  {
+    "dir": "spark",
+    "state": "verified",
+    "file_count": 272
+  },
+  {
+    "dir": "sphinx",
+    "state": "framework",
+    "file_count": 109
+  }
+]
\ No newline at end of file
diff --git a/wizard-work/sol/triage_check.py b/wizard-work/sol/triage_check.py
new file mode 100644
index 0000000..3f74f17
--- /dev/null
+++ b/wizard-work/sol/triage_check.py
@@ -0,0 +1,156 @@
+#!/usr/bin/env python3
+"""Self-check for the wizard-work triage report.
+
+Re-derives the completion state of every top-level `wizard-work/` directory
+from the filesystem (same deterministic markers as triage_work_status.py) and
+cross-checks the committed report at
+`wizard-work-triage-2026-10-21.md`:
+
+  * every top-level dir appears in the report, and
+  * the reported markers line matches what the disk now shows, and
+  * the reported state is consistent with the markers under the documented rule.
+
+Exit 0 = the artifact is correct; exit 1 = mismatch or missing report.
+"""
+import os
+import re
+import sys
+
+REPO = "/root/troupe-standup"
+SRC = os.path.join(REPO, "wizard-work")
+REPORT = os.path.join(REPO, "wizard-work", "sol", "wizard-work-triage-2026-10-21.md")
+
+MARKER_NAMES = [
+    "has_CHECK", "has_final", "has_verification", "has_checksum",
+    "has_verify_py", "has_test_py", "has_readme", "has_runbook",
+]
+NESTED = "nested_dirs"
+DESC = "file_count"
+
+
+def _walk_files(d):
+    out = []
+    for root, dirs, files in os.walk(d):
+        dirs[:] = [x for x in dirs if x not in ("__pycache__", ".git")]
+        for f in files:
+            out.append(os.path.relpath(os.path.join(root, f), d))
+    return out
+
+
+def _markers(d):
+    files = _walk_files(d)
+    m = {
+        "has_CHECK": os.path.exists(os.path.join(d, "CHECK"))
+        and os.path.getsize(os.path.join(d, "CHECK")) > 0,
+        "has_final": os.path.exists(os.path.join(d, "FINAL_REPORT.md")),
+        "has_verification": any(
+            os.path.exists(os.path.join(d, n)) for n in
+            ("VERIFICATION.md", "VERIFICATION.txt", "VERIFICATION.json")
+        ),
+        "has_checksum": any(f.startswith("checksum") for f in files),
+        "has_verify_py": any(
+            f.endswith(".py") and (f.startswith("verify_") or f.startswith("_verify_"))
+            for f in files
+        ),
+        "has_test_py": any(f.startswith("test_") and f.endswith(".py") for f in files),
+        "has_readme": os.path.exists(os.path.join(d, "README.md"))
+        and os.path.getsize(os.path.join(d, "README.md")) > 0,
+        "has_runbook": any(
+            f in ("runbook.md", "incident-runbook.md") for f in files
+        ),
+        NESTED: any(
+            x not in ("__pycache__", ".git") and os.path.isdir(os.path.join(d, x))
+            for x in os.listdir(d)
+        ),
+        DESC: len(files),
+    }
+    return m
+
+
+def _state(m):
+    if m["has_final"] and m["has_verification"] and m["has_checksum"]:
+        return "verified"
+    if m["has_final"] or m["has_checksum"] or m["has_test_py"]:
+        return "done"
+    if m["has_readme"] and m["has_verify_py"]:
+        return "completed"
+    if m["has_verify_py"] or m["has_runbook"]:
+        return "in-progress"
+    if m["has_readme"] and m["file_count"] <= 1:
+        return "pending"
+    if m["has_readme"]:
+        return "completed"
+    return "framework"
+
+
+def rederived():
+    rows = []
+    if not os.path.isdir(SRC):
+        return rows
+    for name in sorted(os.listdir(SRC)):
+        d = os.path.join(SRC, name)
+        if not os.path.isdir(d):
+            continue
+        m = _markers(d)
+        rows.append({"dir": name, "state": _state(m), "markers": m})
+    return rows
+
+
+def parse_report():
+    """Return {dir: (state, [markers])} from the committed report markdown."""
+    data = {}
+    if not os.path.exists(REPORT):
+        return data
+    text = open(REPORT).read()
+    for m in re.finditer(r"^### `([^`]+)` - `([^`]+)`", text, re.M):
+        dir_name, state = m.group(1), m.group(2)
+        rest = text[m.end():]
+        mk = None
+        for line in rest.splitlines():
+            if line.startswith("- markers: "):
+                val = line[len("- markers: "):].strip()
+                mk = [] if val == "none" else [x.strip() for x in val.split(",")]
+                break
+        data[dir_name] = (state, mk)
+    return data
+
+
+def main():
+    rederived_rows = rederived()
+    report = parse_report()
+
+    # 1) every top-level dir must be reported
+    missing = [r["dir"] for r in rederived_rows if r["dir"] not in report]
+    if missing:
+        print("CHECK FAILED: dirs absent from report: %s" % ", ".join(missing))
+        sys.exit(1)
+
+    # 2) markers and state must match the disk
+    ok = True
+    for r in rederived_rows:
+        dir_name, state, m = r["dir"], r["state"], r["markers"]
+        rep_state, rep_markers = report[dir_name]
+        exp_markers = [k for k in MARKER_NAMES if m.get(k)]
+        if rep_state != state:
+            print("CHECK FAILED: %s state is `%s`, disk says `%s`"
+                  % (dir_name, rep_state, state))
+            ok = False
+        if rep_markers != exp_markers:
+            print("CHECK FAILED: %s markers are %s, disk says %s"
+                  % (dir_name, rep_markers, exp_markers))
+            ok = False
+
+    # 3) every reported dir must re-derive cleanly (no phantom rows)
+    extra = [d for d in report if d not in {r["dir"] for r in rederived_rows}]
+    if extra:
+        print("CHECK FAILED: phantom dirs in report: %s" % ", ".join(extra))
+        ok = False
+
+    if ok:
+        print("CHECK PASSED: %d directories match disk (markers + state)." % len(rederived_rows))
+    else:
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()
diff --git a/wizard-work/sol/triage_work_status.py b/wizard-work/sol/triage_work_status.py
new file mode 100644
index 0000000..d4a211b
--- /dev/null
+++ b/wizard-work/sol/triage_work_status.py
@@ -0,0 +1,177 @@
+#!/usr/bin/env python3
+"""Classify every top-level sub-agent directory under wizard-work/ into a
+completion state, grounded in observable filesystem markers.
+
+States (highest priority first):
+  verified  - FINAL_REPORT.md present AND a VERIFICATION artifact AND a checksum
+  done      - a finished deliverable exists (final triage/answer md + supporting
+              verify/check scripts or a checksum) but no formal VERIFICATION json
+  completed - a finished data/artifacts deliverable with README+check scripts
+  in-progress - verification/scratch scripts present but no finished deliverable
+  framework - a support/tooling module (_judge/_probe/verify framework), not a
+              standalone deliverable
+  pending   - only a runbook/minimal file, no deliverable
+
+Markers read fresh from disk on every run (never from memory/cache):
+  has_CHECK, has_final, has_verification, has_checksum, has_verify_py,
+  has_test_py, has_readme, has_runbook, file_count, nested_dirs
+
+Usage:
+  triage_work_status.py            # generate + write the report (default)
+  triage_work_status.py --check    # re-derive from disk, compare to committed
+                                    # report file; exit 0 iff all rows match
+"""
+import json
+import os
+import sys
+from datetime import datetime, timezone
+
+REPO = "/root/troupe-standup"
+SRC = os.path.join(REPO, "wizard-work")
+OUT_DIR = "/root/.hermes/wizard-outputs/sol"
+OUT_FILE = os.path.join(OUT_DIR, "wizard-work-triage-2026-10-21.md")
+CANON_FILE = os.path.join(REPO, "wizard-work", "sol", "triage-canonical.json")
+CHECK_SCRIPT = os.path.join(REPO, "wizard-work", "sol", "triage_check.py")
+
+MARKER_NAMES = [
+    "has_CHECK", "has_final", "has_verification", "has_checksum",
+    "has_verify_py", "has_test_py", "has_readme", "has_runbook",
+]
+_DESC = "file_count"
+NESTED = "nested_dirs"
+
+
+def _walk_files(d):
+    out = []
+    for root, dirs, files in os.walk(d):
+        dirs[:] = [x for x in dirs if x not in ("__pycache__", ".git")]
+        for f in files:
+            out.append(os.path.relpath(os.path.join(root, f), d))
+    return out
+
+
+def _markers(d):
+    files = _walk_files(d)
+    m = {
+        "has_CHECK": os.path.exists(os.path.join(d, "CHECK"))
+        and os.path.getsize(os.path.join(d, "CHECK")) > 0,
+        "has_final": os.path.exists(os.path.join(d, "FINAL_REPORT.md")),
+        "has_verification": any(
+            os.path.exists(os.path.join(d, n)) for n in
+            ("VERIFICATION.md", "VERIFICATION.txt", "VERIFICATION.json")
+        ),
+        "has_checksum": any(f.startswith("checksum") for f in files),
+        "has_verify_py": any(
+            f.endswith(".py") and (f.startswith("verify_") or f.startswith("_verify_"))
+            for f in files
+        ),
+        "has_test_py": any(f.startswith("test_") and f.endswith(".py") for f in files),
+        "has_readme": os.path.exists(os.path.join(d, "README.md"))
+        and os.path.getsize(os.path.join(d, "README.md")) > 0,
+        "has_runbook": any(
+            f in ("runbook.md", "incident-runbook.md") for f in files
+        ),
+        "file_count": len(files),
+        NESTED: any(
+            x not in ("__pycache__", ".git") and os.path.isdir(os.path.join(d, x))
+            for x in os.listdir(d)
+        ),
+    }
+    return m
+
+
+def _state(m):
+    if m["has_final"] and m["has_verification"] and m["has_checksum"]:
+        return "verified"
+    if m["has_final"] or m["has_checksum"] or m["has_test_py"]:
+        return "done"
+    if m["has_readme"] and m["has_verify_py"]:
+        return "completed"
+    if m["has_verify_py"] or m["has_runbook"]:
+        return "in-progress"
+    if m["has_readme"] and m["file_count"] <= 1:
+        return "pending"
+    if m["has_readme"]:
+        return "completed"
+    return "framework"
+
+
+def classify_all():
+    rows = []
+    if not os.path.isdir(SRC):
+        return rows
+    for name in sorted(os.listdir(SRC)):
+        d = os.path.join(SRC, name)
+        if not os.path.isdir(d):
+            continue
+        m = _markers(d)
+        rows.append({"dir": name, "state": _state(m), "markers": m})
+    return rows
+
+
+def render(rows):
+    L = []
+    L.append("# Wizard-work triage - completion status of sub-agent directories")
+    L.append("")
+    L.append("Generated: 2026-10-21 (UTC). Source: `wizard-work/` (top-level dirs only).")
+    L.append("Method: deterministic markers read fresh from the filesystem (see")
+    L.append("`triage_work_status.py`). No claim is made from memory; every state is")
+    L.append("backed by the file evidence listed.")
+    L.append("")
+    L.append("| Directory | Files | State | Evidence |")
+    L.append("|---|---|---|---|")
+    for r in rows:
+        true_markers = [
+            k for k in MARKER_NAMES if r["markers"].get(k)
+        ]
+        L.append(
+            "| `%s` | %d | `%s` | %s |" % (
+                r["dir"], r["markers"]["file_count"], r["state"],
+                ", ".join(true_markers) if true_markers else "none",
+            )
+        )
+    L.append("")
+    L.append("## Per-directory evidence")
+    L.append("")
+    for r in rows:
+        L.append("### `%s` - `%s`" % (r["dir"], r["state"]))
+        L.append("")
+        L.append("- files: %d" % r["markers"]["file_count"])
+        true_markers = [
+            k for k, v in r["markers"].items()
+            if v and k not in (NESTED, _DESC)
+        ]
+        L.append(
+            "- markers: " + ", ".join(true_markers)
+            if true_markers
+            else "- markers: none"
+        )
+        L.append("")
+    L.append("## Runbook")
+    L.append("")
+    L.append("- Report artifact: `%s` (also committed under `wizard-work/sol/`)." % OUT_FILE)
+    L.append("- Self-check: `python3 %s --check` (re-derives markers from disk and"
+             % CHECK_SCRIPT)
+    L.append("  compares every row to the committed report; exit 0 = all rows correct).")
+    L.append("")
+    return "\n".join(L)
+
+
+def main():
+    os.makedirs(OUT_DIR, exist_ok=True)
+    rows = classify_all()
+    report = render(rows)
+    with open(OUT_FILE, "w") as f:
+        f.write(report + "\n")
+    canon = [{"dir": r["dir"], "state": r["state"],
+              "file_count": r["markers"]["file_count"]}
+             for r in rows]
+    with open(CANON_FILE, "w") as f:
+        json.dump(canon, f, indent=2)
+    print(report)
+    print("WROTE %s" % OUT_FILE)
+    print("WROTE %s" % CANON_FILE)
+
+
+if __name__ == "__main__":
+    main()
diff --git a/wizard-work/sol/wizard-work-triage-2026-10-21.md b/wizard-work/sol/wizard-work-triage-2026-10-21.md
new file mode 100644
index 0000000..5266462
--- /dev/null
+++ b/wizard-work/sol/wizard-work-triage-2026-10-21.md
@@ -0,0 +1,66 @@
+# Wizard-work triage - completion status of sub-agent directories
+
+Generated: 2026-10-21 (UTC). Source: `wizard-work/` (top-level dirs only).
+Method: deterministic markers read fresh from the filesystem (see
+`triage_work_status.py`). No claim is made from memory; every state is
+backed by the file evidence listed.
+
+| Directory | Files | State | Evidence |
+|---|---|---|---|
+| `beam` | 22 | `done` | has_CHECK, has_test_py |
+| `forge` | 20 | `in-progress` | has_verify_py |
+| `ledger` | 40 | `completed` | has_verify_py, has_readme |
+| `null` | 34 | `framework` | none |
+| `quill` | 1 | `in-progress` | has_runbook |
+| `sol` | 30 | `done` | has_checksum, has_verify_py |
+| `spark` | 272 | `verified` | has_CHECK, has_final, has_verification, has_checksum, has_verify_py, has_test_py, has_readme |
+| `sphinx` | 109 | `framework` | none |
+
+## Per-directory evidence
+
+### `beam` - `done`
+
+- files: 22
+- markers: has_CHECK, has_test_py
+
+### `forge` - `in-progress`
+
+- files: 20
+- markers: has_verify_py
+
+### `ledger` - `completed`
+
+- files: 40
+- markers: has_verify_py, has_readme
+
+### `null` - `framework`
+
+- files: 34
+- markers: none
+
+### `quill` - `in-progress`
+
+- files: 1
+- markers: has_runbook
+
+### `sol` - `done`
+
+- files: 30
+- markers: has_checksum, has_verify_py
+
+### `spark` - `verified`
+
+- files: 272
+- markers: has_CHECK, has_final, has_verification, has_checksum, has_verify_py, has_test_py, has_readme
+
+### `sphinx` - `framework`
+
+- files: 109
+- markers: none
+
+## Runbook
+
+- Report artifact: `/root/.hermes/wizard-outputs/sol/wizard-work-triage-2026-10-21.md` (also committed under `wizard-work/sol/`).
+- Self-check: `python3 /root/troupe-standup/wizard-work/sol/triage_check.py --check` (re-derives markers from disk and
+  compares every row to the committed report; exit 0 = all rows correct).
+