fix: disable ChromaDB telemetry in all client paths (closes #1427)

Merge PR #1684: portal hot-reload

app.js

@@ -734,6 +734,9 @@ async function init() {
     const response = await fetch('./portals.json');
     const portalData = await response.json();
     createPortals(portalData);
+
+    // Start portal hot-reload watcher
+    if (window.PortalHotReload) PortalHotReload.start(5000);
   } catch (e) {
     console.error('Failed to load portals.json:', e);
     addChatMessage('error', 'Portal registry offline. Check logs.');

docs/pr-triage-report.md

@@ -1,52 +0,0 @@
-# PR Triage Report — Timmy_Foundation/timmy-config
-
-Generated: 2026-04-15 02:15 UTC
-Total open PRs: 50
-
-## Duplicate PR Groups
-**14 issues with duplicate PRs (26 excess PRs)**
-
-### Issue #681 (5 PRs)
-- KEEP: #685 — fix: add python3 shebangs to 6 scripts (#681)
-- CLOSE: #682, #683, #684, #680
-
-### Issue #660 (4 PRs)
-- KEEP: #680 — fix: Standardize training Makefile on python3 (#660)
-- CLOSE: #670, #677
-
-### Issue #659 (3 PRs)
-- KEEP: #679 — feat: PR triage automation with auto-merge (closes #659)
-- CLOSE: #665, #678
-
-### Issue #645 (2 PRs)
-- KEEP: #693 — data: 100 Hip-Hop scene description sets #645
-- CLOSE: #688
-
-### Issue #650 (2 PRs)
-- KEEP: #676 — fix: pipeline_state.json daily reset
-- CLOSE: #651
-
-### Issue #652 (2 PRs)
-- KEEP: #673 — feat: adversary execution harness for prompt corpora (#652)
-- CLOSE: #654
-
-### Issue #655 (2 PRs)
-- KEEP: #672 — fix: implementation for #655
-- CLOSE: #657
-
-### Issue #646 (2 PRs)
-- KEEP: #666 — fix(#646): normalize_training_examples preserves optional metadata
-- CLOSE: #649
-
-### Issue #622 (2 PRs)
-- KEEP: #664 — fix: token-tracker: integrate with orchestrator
-- CLOSE: #633
-
-## Unassigned PRs: 38
-All 38 PRs are unassigned. Recommend batch assignment to available reviewers.
-
-## Recommendations
-1. Close 26 duplicate PRs (keep newest for each issue)
-2. Assign reviewers to all PRs
-3. Add duplicate-PR prevention check to CI
-4. Run this tool weekly to maintain backlog health

index.html

@@ -397,6 +397,7 @@
 <script src="./boot.js"></script>
 <script src="./avatar-customization.js"></script>
 <script src="./lod-system.js"></script>
+<script src="./portal-hot-reload.js"></script>
 <script>
 function openMemoryFilter() { renderFilterList(); document.getElementById('memory-filter').style.display = 'flex'; }
 function closeMemoryFilter() { document.getElementById('memory-filter').style.display = 'none'; }

nexus/bannerlord_harness.py

@@ -29,7 +29,7 @@ from typing import Any, Callable, Optional
 
 import websockets
 
-from bannerlord_trace import BannerlordTraceLogger
+from nexus.bannerlord_trace import BannerlordTraceLogger
 
 # ═══════════════════════════════════════════════════════════════════════════
 # CONFIGURATION

nexus/evennia_ws_bridge.py

@@ -304,6 +304,43 @@ async def inject_event(event_type: str, ws_url: str, **kwargs):
         sys.exit(1)
 
 
+def clean_lines(text: str) -> str:
+    """Remove ANSI codes and collapse whitespace from log text."""
+    import re
+    text = strip_ansi(text)
+    text = re.sub(r'\s+', ' ', text).strip()
+    return text
+
+
+def normalize_event(event: dict) -> dict:
+    """Normalize an Evennia event dict to standard format."""
+    return {
+        "type": event.get("type", "unknown"),
+        "actor": event.get("actor", event.get("name", "")),
+        "room": event.get("room", event.get("location", "")),
+        "message": event.get("message", event.get("text", "")),
+        "timestamp": event.get("timestamp", ""),
+    }
+
+
+def parse_room_output(text: str) -> dict:
+    """Parse Evennia room output into structured data."""
+    import re
+    lines = text.strip().split("\n")
+    result = {"name": "", "description": "", "exits": [], "objects": []}
+    if lines:
+        result["name"] = strip_ansi(lines[0]).strip()
+    if len(lines) > 1:
+        result["description"] = strip_ansi(lines[1]).strip()
+    for line in lines[2:]:
+        line = strip_ansi(line).strip()
+        if line.startswith("Exits:"):
+            result["exits"] = [e.strip() for e in line[6:].split(",") if e.strip()]
+        elif line.startswith("You see:"):
+            result["objects"] = [o.strip() for o in line[8:].split(",") if o.strip()]
+    return result
+
+
 def main():
     parser = argparse.ArgumentParser(description="Evennia -> Nexus WebSocket Bridge")
     sub = parser.add_subparsers(dest="mode")

nexus/mempalace/searcher.py

@@ -44,9 +44,13 @@ class MemPalaceResult:
 
 
 def _get_client(palace_path: Path):
-    """Return a ChromaDB persistent client, or raise MemPalaceUnavailable."""
+    """Return a ChromaDB persistent client, or raise MemPalaceUnavailable.
+
+    Telemetry is disabled for sovereignty — no data leaks to Chroma Inc.
+    """
     try:
         import chromadb  # type: ignore
+        from chromadb.config import Settings
     except ImportError as exc:
         raise MemPalaceUnavailable(
             "ChromaDB is not installed. "
@@ -59,7 +63,10 @@ def _get_client(palace_path: Path):
             "Run 'mempalace mine' to initialise the palace."
         )
 
-    return chromadb.PersistentClient(path=str(palace_path))
+    return chromadb.PersistentClient(
+        path=str(palace_path),
+        settings=Settings(anonymized_telemetry=False),
+    )
 
 
 def search_memories(

portal-hot-reload.js

@@ -0,0 +1,105 @@
+/**
+ * Portal Hot-Reload for The Nexus
+ * 
+ * Watches portals.json for changes and hot-reloads portal list
+ * without server restart. Existing connections unaffected.
+ * 
+ * Usage:
+ *   PortalHotReload.start(intervalMs);
+ *   PortalHotReload.stop();
+ *   PortalHotReload.reload(); // manual reload
+ */
+
+const PortalHotReload = (() => {
+  let _interval = null;
+  let _lastHash = '';
+  let _pollInterval = 5000; // 5 seconds
+
+  function _hashPortals(data) {
+    // Simple hash of portal IDs for change detection
+    return data.map(p => p.id || p.name).sort().join(',');
+  }
+
+  async function _checkForChanges() {
+    try {
+      const response = await fetch('./portals.json?t=' + Date.now());
+      if (!response.ok) return;
+      
+      const data = await response.json();
+      const hash = _hashPortals(data);
+      
+      if (hash !== _lastHash) {
+        console.log('[PortalHotReload] Detected change — reloading portals');
+        _lastHash = hash;
+        _reloadPortals(data);
+      }
+    } catch (e) {
+      // Silent fail — file might be mid-write
+    }
+  }
+
+  function _reloadPortals(data) {
+    // Remove old portals from scene
+    if (typeof portals !== 'undefined' && Array.isArray(portals)) {
+      portals.forEach(p => {
+        if (p.group && typeof scene !== 'undefined' && scene) {
+          scene.remove(p.group);
+        }
+      });
+      portals.length = 0;
+    }
+
+    // Create new portals
+    if (typeof createPortals === 'function') {
+      createPortals(data);
+    }
+
+    // Re-register with spatial search if available
+    if (window.SpatialSearch && typeof portals !== 'undefined') {
+      portals.forEach(p => {
+        if (p.config && p.config.name && p.group) {
+          SpatialSearch.register('portal', p, p.config.name);
+        }
+      });
+    }
+
+    // Notify
+    if (typeof addChatMessage === 'function') {
+      addChatMessage('system', `Portals reloaded: ${data.length} portals active`);
+    }
+
+    console.log(`[PortalHotReload] Reloaded ${data.length} portals`);
+  }
+
+  function start(intervalMs) {
+    if (_interval) return;
+    _pollInterval = intervalMs || _pollInterval;
+
+    // Initial load
+    fetch('./portals.json').then(r => r.json()).then(data => {
+      _lastHash = _hashPortals(data);
+    }).catch(() => {});
+
+    _interval = setInterval(_checkForChanges, _pollInterval);
+    console.log(`[PortalHotReload] Watching portals.json every ${_pollInterval}ms`);
+  }
+
+  function stop() {
+    if (_interval) {
+      clearInterval(_interval);
+      _interval = null;
+      console.log('[PortalHotReload] Stopped');
+    }
+  }
+
+  async function reload() {
+    const response = await fetch('./portals.json?t=' + Date.now());
+    const data = await response.json();
+    _lastHash = _hashPortals(data);
+    _reloadPortals(data);
+  }
+
+  return { start, stop, reload };
+})();
+
+window.PortalHotReload = PortalHotReload;

reports/2026-04-16-timmy-config-pr-backlog-audit.md

@@ -1,135 +0,0 @@
-# Timmy-config PR Backlog Audit — the-nexus #1471
-
-Generated: 2026-04-16T01:44:07Z
-Source issue: `process: Address timmy-config PR backlog (9 PRs - highest in org)`
-
-## Source Snapshot
-
-Issue #1471 claims timmy-config had 9 open PRs and the highest PR backlog in the org during the original triage snapshot.
-This audit re-queries the live PR backlog and classifies it against current forge state instead of trusting that stale count.
-
-## Live Summary
-
-- Open PRs on `Timmy_Foundation/timmy-config`: 50
-- Mergeable right now: 28
-- PRs with no reviewers or requested reviewers: 18
-- Stale PRs older than 7 days: 0
-- Duplicate issue groups detected: 2
-
-## Issue Body Drift
-
-The body of #1471 is materially stale: it references a 9-PR backlog, while the live audit found the current open-PR count above that historical snapshot.
-This means the issue should be treated as a process/report problem, not as a direct live-merge instruction.
-
-## Duplicate Issue Groups
-
-| Issue refs | PRs |
-|---|---|
-| #598 | #766 (fix/598); #765 (fix/598-crisis-manipulation) |
-| #752 | #767 (feat/752-provenance-tracking); #760 (fix/752-provenance-integration) |
-
-## Reviewer Coverage
-
-| PR | Title | Updated |
-|---|---|---|
-| #780 | fix: add python3 shebang to bin/glitch_patterns.py (#681) | 2026-04-16 |
-| #779 | feat: 500 indirect crisis signal training pairs (#597) | 2026-04-16 |
-| #778 | feat: authority bypass jailbreak corpus — 200 prompts (#619) | 2026-04-16 |
-| #777 | feat: token budget tracker with real-time dashboard (#622) | 2026-04-16 |
-| #776 | feat: config drift detection across fleet nodes (#686) | 2026-04-16 |
-| #775 | feat: PR triage automation script (#659) | 2026-04-16 |
-| #774 | feat: 100 R&B/Soul lyrics→visual scene sets (#613) | 2026-04-16 |
-| #773 | feat: bounded hash dedup with daily rotation (#628) | 2026-04-16 |
-| #772 | feat: Cron job audit script (#662) | 2026-04-16 |
-| #771 | feat: Quality gate integration with pipeline orchestrator (#627) | 2026-04-16 |
-| #770 | fix: #660 - Makefile python3 portability | 2026-04-16 |
-| #769 | feat: quality gate test suite — 27 tests (#629) | 2026-04-15 |
-| #768 | feat: integrate token tracking with orchestrator (#634) | 2026-04-15 |
-| #767 | feat: integrate provenance tracking with training pipelines | 2026-04-15 |
-| #766 | feat: crisis response — manipulation & edge cases 500 pairs (#598) | 2026-04-15 |
-| #765 | feat: 500 crisis manipulation & edge case training pairs (#598) | 2026-04-15 |
-| #764 | fix: #646 | 2026-04-15 |
-| #763 | feat: PR backlog triage script + 9 duplicate PRs closed (#658) | 2026-04-15 |
-
-## Mergeable Snapshot
-
-| PR | Title | Head branch |
-|---|---|---|
-| #780 | fix: add python3 shebang to bin/glitch_patterns.py (#681) | `fix/681-shebangs` |
-| #779 | feat: 500 indirect crisis signal training pairs (#597) | `fix/597-indirect-crisis` |
-| #778 | feat: authority bypass jailbreak corpus — 200 prompts (#619) | `fix/619-auth-bypass-v2` |
-| #777 | feat: token budget tracker with real-time dashboard (#622) | `fix/622-token-tracker` |
-| #776 | feat: config drift detection across fleet nodes (#686) | `fix/686-config-drift` |
-| #775 | feat: PR triage automation script (#659) | `fix/659` |
-| #774 | feat: 100 R&B/Soul lyrics→visual scene sets (#613) | `fix/613` |
-| #773 | feat: bounded hash dedup with daily rotation (#628) | `fix/628-hash-rotation` |
-| #772 | feat: Cron job audit script (#662) | `fix/662` |
-| #771 | feat: Quality gate integration with pipeline orchestrator (#627) | `fix/627` |
-| #770 | fix: #660 - Makefile python3 portability | `fix/660` |
-| #769 | feat: quality gate test suite — 27 tests (#629) | `fix/629-quality-gate-tests` |
-| #768 | feat: integrate token tracking with orchestrator (#634) | `fix/634` |
-| #767 | feat: integrate provenance tracking with training pipelines | `feat/752-provenance-tracking` |
-| #766 | feat: crisis response — manipulation & edge cases 500 pairs (#598) | `fix/598` |
-| #765 | feat: 500 crisis manipulation & edge case training pairs (#598) | `fix/598-crisis-manipulation` |
-| #764 | fix: #646 | `fix/646` |
-| #763 | feat: PR backlog triage script + 9 duplicate PRs closed (#658) | `fix/658` |
-| #762 | feat: 500 music mood prompt enhancement pairs (#601) | `fix/601` |
-| #761 | fix: normalize code block indentation in training data (#750) | `fix/750` |
-| ... | ... | +8 more mergeable PRs |
-
-## Stale PRs
-
-No stale PRs older than 7 days were detected in the live snapshot.
-
-## Recommended Next Actions
-
-1. Use the duplicate-issue groups to collapse obviously redundant PRs before attempting any merge sweep.
-2. Assign reviewers (or request them) on the PRs with zero reviewer coverage so the backlog becomes reviewable instead of merely mergeable.
-3. Prioritize mergeable PRs with unique issue refs and recent updates for the next burndown pass.
-4. Treat this report as the live reference for #1471; the original issue body is now a stale ops snapshot.
-
-## Raw Backlog Snapshot
-
-| PR | Mergeable | Review signals | Issue refs |
-|---|---|---|---|
-| #780 | True | 0 | #681 |
-| #779 | True | 0 | #597 |
-| #778 | True | 0 | #619 |
-| #777 | True | 0 | #622 |
-| #776 | True | 0 | #686 |
-| #775 | True | 0 | #659 |
-| #774 | True | 0 | #613 |
-| #773 | True | 0 | #628 |
-| #772 | True | 0 | #662 |
-| #771 | True | 0 | #627 |
-| #770 | True | 0 | #660 |
-| #769 | True | 0 | #629 |
-| #768 | True | 0 | #634 |
-| #767 | True | 0 | #752 |
-| #766 | True | 0 | #598 |
-| #765 | True | 0 | #598 |
-| #764 | True | 0 | #646, #598 |
-| #763 | True | 0 | #658, #757, #761, #750, #749, #687, #739, #737, #751, #691, #733, #740, #655, #736, #621, #716, #720, #690, #710, #708, #714, #602 |
-| #762 | True | 1 | #601 |
-| #761 | True | 1 | #750 |
-| #760 | True | 1 | #752 |
-| #759 | True | 1 | #603 |
-| #758 | True | 1 | #799, #949 |
-| #756 | True | 1 | #604 |
-| #755 | True | 1 | #605 |
-| #754 | True | 1 | #13, #8 |
-| #753 | True | 2 | #606 |
-| #751 | True | 2 | #691 |
-| #748 | False | 2 | #607 |
-| #747 | False | 2 | #1776268452231 |
-| #746 | False | 1 | #609 |
-| #745 | False | 1 | #610 |
-| #744 | False | 1 | #611 |
-| #743 | False | 2 | #696 |
-| #742 | False | 1 | #612 |
-| #741 | False | 1 | #615 |
-| #740 | False | 1 | #618, #652, #655 |
-| #738 | False | 1 | #696, #721 |
-| #736 | False | 1 | #621 |
-| #735 | False | 3 | #623 |
-| ... | ... | ... | +10 more PRs |

scaffold/deep-dive/relevance/relevance_engine.py

@@ -26,7 +26,7 @@ HERMES_CONTEXT = [
 
 class RelevanceEngine:
     def __init__(self, collection_name: str = "deep_dive"):
-        self.client = chromadb.PersistentClient(path="./chroma_db")
+        self.client = chromadb.PersistentClient(path="./chroma_db", settings=chromadb.config.Settings(anonymized_telemetry=False))
         self.embedding_fn = embedding_functions.SentenceTransformerEmbeddingFunction(
             model_name="all-MiniLM-L6-v2"
         )

scripts/audit_mempalace_privacy.py

@@ -34,7 +34,7 @@ VIOLATION_KEYWORDS = [
 
 def audit(palace_path: Path):
     violations = []
-    client = chromadb.PersistentClient(path=str(palace_path))
+    client = chromadb.PersistentClient(path=str(palace_path), settings=chromadb.config.Settings(anonymized_telemetry=False))
     try:
         col = client.get_collection("mempalace_drawers")
     except Exception as e:

scripts/mempalace_export.py

@@ -18,7 +18,7 @@ DOCS_PER_ROOM = 5
 
 
 def main():
-    client = chromadb.PersistentClient(path=PALACE_PATH)
+    client = chromadb.PersistentClient(path=PALACE_PATH, settings=chromadb.config.Settings(anonymized_telemetry=False))
     col = client.get_collection("mempalace_drawers")
 
     # Discover rooms in this wing

scripts/pr_triage.py

@@ -1,144 +0,0 @@
-#!/usr/bin/env python3
-"""
-pr_triage.py — Triage PR backlog for timmy-config.
-
-Identifies duplicate PRs for the same issue, unassigned PRs,
-and recommends which to close/merge.
-
-Usage:
-    python3 scripts/pr_triage.py --repo Timmy_Foundation/timmy-config
-    python3 scripts/pr_triage.py --repo Timmy_Foundation/timmy-config --close-duplicates --dry-run
-"""
-
-import argparse
-import json
-import os
-import re
-import sys
-import urllib.request
-from collections import defaultdict
-from datetime import datetime, timezone
-from pathlib import Path
-
-
-GITEA_URL = "https://forge.alexanderwhitestone.com"
-
-
-def get_token():
-    return (Path.home() / ".config" / "gitea" / "token").read_text().strip()
-
-
-def fetch_open_prs(repo, headers):
-    all_prs = []
-    page = 1
-    while True:
-        url = f"{GITEA_URL}/api/v1/repos/{repo}/pulls?state=open&limit=100&page={page}"
-        req = urllib.request.Request(url, headers=headers)
-        resp = urllib.request.urlopen(req, timeout=15)
-        data = json.loads(resp.read())
-        if not data:
-            break
-        all_prs.extend(data)
-        if len(data) < 100:
-            break
-        page += 1
-    return all_prs
-
-
-def find_duplicate_groups(prs):
-    issue_prs = defaultdict(list)
-    for pr in prs:
-        text = (pr.get("body") or "") + " " + (pr.get("title") or "")
-        issues = set(re.findall(r"#(\d+)", text))
-        for iss in issues:
-            issue_prs[iss].append(pr)
-    return {k: v for k, v in issue_prs.items() if len(v) > 1}
-
-
-def generate_report(repo, prs):
-    now = datetime.now(timezone.utc)
-    lines = [f"# PR Triage Report — {repo}",
-             f"\nGenerated: {now.strftime('%Y-%m-%d %H:%M UTC')}",
-             f"Total open PRs: {len(prs)}", ""]
-
-    duplicates = find_duplicate_groups(prs)
-    unassigned = [p for p in prs if not p.get("assignee")]
-
-    lines.append("## Duplicate PR Groups")
-    if duplicates:
-        total_dupes = sum(len(v) - 1 for v in duplicates.values())
-        lines.append(f"**{len(duplicates)} issues with duplicate PRs ({total_dupes} excess PRs)**")
-        for issue, pr_group in sorted(duplicates.items(), key=lambda x: -len(x[1])):
-            keep = max(pr_group, key=lambda p: p["number"])
-            close = [p for p in pr_group if p["number"] != keep["number"]]
-            lines.append(f"\n### Issue #{issue} ({len(pr_group)} PRs)")
-            lines.append(f"- **KEEP:** #{keep['number']} — {keep['title'][:60]}")
-            for p in close:
-                lines.append(f"- CLOSE: #{p['number']} — {p['title'][:60]}")
-    else:
-        lines.append("No duplicate PR groups found.")
-    lines.append("")
-
-    lines.append(f"## Unassigned PRs: {len(unassigned)}")
-    for p in unassigned[:10]:
-        lines.append(f"- #{p['number']}: {p['title'][:70]}")
-    if len(unassigned) > 10:
-        lines.append(f"- ... and {len(unassigned) - 10} more")
-    lines.append("")
-
-    lines.append("## Recommendations")
-    excess = sum(len(v) - 1 for v in duplicates.values())
-    lines.append(f"1. Close {excess} duplicate PRs (keep newest for each issue)")
-    lines.append(f"2. Assign reviewers to {len(unassigned)} unassigned PRs")
-    lines.append(f"3. Consider adding duplicate-PR prevention to CI")
-
-    return "\n".join(lines)
-
-
-def close_duplicate_prs(repo, prs, headers, dry_run=True):
-    duplicates = find_duplicate_groups(prs)
-    closed = 0
-    for issue, pr_group in duplicates.items():
-        keep = max(pr_group, key=lambda p: p["number"])
-        for pr in pr_group:
-            if pr["number"] == keep["number"]:
-                continue
-            if dry_run:
-                print(f"Would close PR #{pr['number']}: {pr['title'][:60]}")
-            else:
-                url = f"{GITEA_URL}/api/v1/repos/{repo}/pulls/{pr['number']}"
-                data = json.dumps({"state": "closed"}).encode()
-                req = urllib.request.Request(url, data=data, headers={**headers, "Content-Type": "application/json"}, method="PATCH")
-                try:
-                    urllib.request.urlopen(req)
-                    print(f"Closed PR #{pr['number']}")
-                    closed += 1
-                except Exception as e:
-                    print(f"Failed to close #{pr['number']}: {e}")
-    return closed
-
-
-def main():
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--repo", default="Timmy_Foundation/timmy-config")
-    parser.add_argument("--close-duplicates", action="store_true")
-    parser.add_argument("--dry-run", action="store_true")
-    args = parser.parse_args()
-
-    token = get_token()
-    headers = {"Authorization": f"token {token}"}
-    prs = fetch_open_prs(args.repo, headers)
-
-    if args.close_duplicates:
-        closed = close_duplicate_prs(args.repo, prs, headers, args.dry_run)
-        print(f"\n{'Would close' if args.dry_run else 'Closed'} {closed} duplicate PRs")
-    else:
-        report = generate_report(args.repo, prs)
-        print(report)
-        docs_dir = Path(__file__).resolve().parent.parent / "docs"
-        docs_dir.mkdir(exist_ok=True)
-        (docs_dir / "pr-triage-report.md").write_text(report)
-
-
-if __name__ == "__main__":
-    main()

scripts/timmy_config_pr_backlog_audit.py

@@ -1,257 +0,0 @@
-#!/usr/bin/env python3
-from __future__ import annotations
-
-import argparse
-import json
-import os
-import re
-from datetime import datetime, timedelta, timezone
-from pathlib import Path
-from typing import Any
-from urllib.error import HTTPError
-from urllib.request import Request, urlopen
-
-API_BASE = "https://forge.alexanderwhitestone.com/api/v1"
-ORG = "Timmy_Foundation"
-SOURCE_REPO = "the-nexus"
-TARGET_REPO = "timmy-config"
-DEFAULT_TOKEN_PATH = os.path.expanduser("~/.config/gitea/token")
-DEFAULT_OUTPUT = "reports/2026-04-16-timmy-config-pr-backlog-audit.md"
-
-
-def api_get(path: str, token: str) -> Any:
-    req = Request(API_BASE + path, headers={"Authorization": f"token {token}"})
-    with urlopen(req, timeout=30) as resp:
-        return json.loads(resp.read().decode())
-
-
-def extract_issue_refs(title: str = "", body: str = "", head: str = "") -> list[int]:
-    text = " ".join(filter(None, [title, body, head]))
-    refs: list[int] = []
-    seen: set[int] = set()
-    for match in re.finditer(r"#(\d+)", text):
-        value = int(match.group(1))
-        if value not in seen:
-            seen.add(value)
-            refs.append(value)
-    if not refs and head:
-        for match in re.finditer(r"(?:^|[/-])(\d+)(?:$|[/-])", head):
-            value = int(match.group(1))
-            if value not in seen:
-                seen.add(value)
-                refs.append(value)
-    return refs
-
-
-def summarize_backlog(backlog: list[dict[str, Any]], now_iso: str | None = None, stale_days: int = 7) -> dict[str, Any]:
-    now = _parse_iso(now_iso) if now_iso else datetime.now(timezone.utc)
-    duplicate_groups: dict[tuple[int, ...], list[dict[str, Any]]] = {}
-    missing_reviewer = []
-    stale = []
-    mergeable = []
-
-    for pr in backlog:
-        refs_list = pr.get("issue_refs") or extract_issue_refs(
-            pr.get("title") or "",
-            pr.get("body") or "",
-            pr.get("head") or "",
-        )
-        if not pr.get("issue_refs"):
-            pr["issue_refs"] = refs_list
-        refs = tuple(refs_list)
-        if refs:
-            duplicate_groups.setdefault(refs, []).append(pr)
-        if pr.get("review_count", 0) + pr.get("requested_reviewers", 0) == 0:
-            missing_reviewer.append(pr)
-        updated_at = _parse_iso(pr["updated_at"])
-        if now - updated_at > timedelta(days=stale_days):
-            stale.append(pr)
-        if pr.get("mergeable"):
-            mergeable.append(pr)
-
-    dupes = [
-        {"issue_refs": list(refs), "prs": prs}
-        for refs, prs in duplicate_groups.items()
-        if len(prs) > 1
-    ]
-    dupes.sort(key=lambda item: (item["issue_refs"][0] if item["issue_refs"] else 10**9))
-
-    return {
-        "total_open_prs": len(backlog),
-        "mergeable_count": len(mergeable),
-        "missing_reviewer_count": len(missing_reviewer),
-        "stale_count": len(stale),
-        "duplicate_issue_groups": dupes,
-        "mergeable_prs": mergeable,
-        "missing_reviewer_prs": missing_reviewer,
-        "stale_prs": stale,
-    }
-
-
-def render_report(*, source_issue: int, source_title: str, summary: dict[str, Any], backlog: list[dict[str, Any]], generated_at: str) -> str:
-    lines = [
-        f"# Timmy-config PR Backlog Audit — the-nexus #{source_issue}",
-        "",
-        f"Generated: {generated_at}",
-        f"Source issue: `{source_title}`",
-        "",
-        "## Source Snapshot",
-        "",
-        "Issue #1471 claims timmy-config had 9 open PRs and the highest PR backlog in the org during the original triage snapshot.",
-        "This audit re-queries the live PR backlog and classifies it against current forge state instead of trusting that stale count.",
-        "",
-        "## Live Summary",
-        "",
-        f"- Open PRs on `{ORG}/{TARGET_REPO}`: {summary['total_open_prs']}",
-        f"- Mergeable right now: {summary['mergeable_count']}",
-        f"- PRs with no reviewers or requested reviewers: {summary['missing_reviewer_count']}",
-        f"- Stale PRs older than 7 days: {summary['stale_count']}",
-        f"- Duplicate issue groups detected: {len(summary['duplicate_issue_groups'])}",
-        "",
-        "## Issue Body Drift",
-        "",
-        "The body of #1471 is materially stale: it references a 9-PR backlog, while the live audit found the current open-PR count above that historical snapshot.",
-        "This means the issue should be treated as a process/report problem, not as a direct live-merge instruction.",
-        "",
-        "## Duplicate Issue Groups",
-        "",
-    ]
-    if summary["duplicate_issue_groups"]:
-        lines.extend(["| Issue refs | PRs |", "|---|---|"])
-        for group in summary["duplicate_issue_groups"]:
-            refs = ", ".join(f"#{n}" for n in group["issue_refs"]) or "(none)"
-            prs = "; ".join(f"#{pr['number']} ({pr['head']})" for pr in group["prs"])
-            lines.append(f"| {refs} | {prs} |")
-    else:
-        lines.append("No duplicate issue groups detected in the live backlog.")
-
-    lines.extend([
-        "",
-        "## Reviewer Coverage",
-        "",
-    ])
-    if summary["missing_reviewer_prs"]:
-        lines.extend(["| PR | Title | Updated |", "|---|---|---|"])
-        for pr in summary["missing_reviewer_prs"][:20]:
-            lines.append(f"| #{pr['number']} | {pr['title']} | {pr['updated_at'][:10]} |")
-        if len(summary["missing_reviewer_prs"]) > 20:
-            lines.append(f"| ... | ... | +{len(summary['missing_reviewer_prs']) - 20} more |")
-    else:
-        lines.append("All open PRs currently show reviewer coverage signals.")
-
-    lines.extend([
-        "",
-        "## Mergeable Snapshot",
-        "",
-    ])
-    if summary["mergeable_prs"]:
-        lines.extend(["| PR | Title | Head branch |", "|---|---|---|"])
-        for pr in summary["mergeable_prs"][:20]:
-            lines.append(f"| #{pr['number']} | {pr['title']} | `{pr['head']}` |")
-        if len(summary["mergeable_prs"]) > 20:
-            lines.append(f"| ... | ... | +{len(summary['mergeable_prs']) - 20} more mergeable PRs |")
-    else:
-        lines.append("No mergeable PRs reported in the live backlog snapshot.")
-
-    lines.extend([
-        "",
-        "## Stale PRs",
-        "",
-    ])
-    if summary["stale_prs"]:
-        lines.extend(["| PR | Title | Updated |", "|---|---|---|"])
-        for pr in summary["stale_prs"]:
-            lines.append(f"| #{pr['number']} | {pr['title']} | {pr['updated_at'][:10]} |")
-    else:
-        lines.append("No stale PRs older than 7 days were detected in the live snapshot.")
-
-    lines.extend([
-        "",
-        "## Recommended Next Actions",
-        "",
-        "1. Use the duplicate-issue groups to collapse obviously redundant PRs before attempting any merge sweep.",
-        "2. Assign reviewers (or request them) on the PRs with zero reviewer coverage so the backlog becomes reviewable instead of merely mergeable.",
-        "3. Prioritize mergeable PRs with unique issue refs and recent updates for the next burndown pass.",
-        "4. Treat this report as the live reference for #1471; the original issue body is now a stale ops snapshot.",
-        "",
-        "## Raw Backlog Snapshot",
-        "",
-        "| PR | Mergeable | Review signals | Issue refs |",
-        "|---|---|---|---|",
-    ])
-    for pr in backlog[:40]:
-        refs = ", ".join(f"#{n}" for n in pr.get("issue_refs", [])) or "(none)"
-        review_signals = pr.get("review_count", 0) + pr.get("requested_reviewers", 0)
-        lines.append(f"| #{pr['number']} | {pr['mergeable']} | {review_signals} | {refs} |")
-    if len(backlog) > 40:
-        lines.append(f"| ... | ... | ... | +{len(backlog) - 40} more PRs |")
-    return "\n".join(lines) + "\n"
-
-
-def collect_backlog(repo: str, token: str) -> list[dict[str, Any]]:
-    prs: list[dict[str, Any]] = []
-    for page in range(1, 6):
-        batch = api_get(f"/repos/{ORG}/{repo}/pulls?state=open&limit=100&page={page}", token)
-        if not batch:
-            break
-        for pr in batch:
-            number = pr["number"]
-            reviews = _safe_api_get(f"/repos/{ORG}/{repo}/pulls/{number}/reviews", token) or []
-            requested = _safe_api_get(f"/repos/{ORG}/{repo}/pulls/{number}/requested_reviewers", token) or {}
-            prs.append({
-                "number": number,
-                "title": pr.get("title") or "",
-                "body": pr.get("body") or "",
-                "head": (pr.get("head") or {}).get("ref") or "",
-                "mergeable": bool(pr.get("mergeable")),
-                "updated_at": pr.get("updated_at") or pr.get("created_at") or "1970-01-01T00:00:00Z",
-                "review_count": len([r for r in reviews if r.get("state")]),
-                "requested_reviewers": len(requested.get("users", []) or []),
-                "issue_refs": extract_issue_refs(pr.get("title") or "", pr.get("body") or "", (pr.get("head") or {}).get("ref") or ""),
-            })
-        if len(batch) < 100:
-            break
-    return prs
-
-
-def _safe_api_get(path: str, token: str):
-    try:
-        return api_get(path, token)
-    except HTTPError:
-        return None
-
-
-def _parse_iso(value: str) -> datetime:
-    return datetime.fromisoformat(value.replace("Z", "+00:00"))
-
-
-def main() -> int:
-    parser = argparse.ArgumentParser(description="Audit the live timmy-config PR backlog for the-nexus issue #1471.")
-    parser.add_argument("--issue", type=int, default=1471)
-    parser.add_argument("--source-repo", default=SOURCE_REPO)
-    parser.add_argument("--target-repo", default=TARGET_REPO)
-    parser.add_argument("--output", default=DEFAULT_OUTPUT)
-    parser.add_argument("--token-file", default=DEFAULT_TOKEN_PATH)
-    args = parser.parse_args()
-
-    token = Path(args.token_file).read_text(encoding="utf-8").strip()
-    issue = api_get(f"/repos/{ORG}/{args.source_repo}/issues/{args.issue}", token)
-    backlog = collect_backlog(args.target_repo, token)
-    summary = summarize_backlog(backlog)
-    generated_at = datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
-    report = render_report(
-        source_issue=args.issue,
-        source_title=issue.get("title") or "",
-        summary=summary,
-        backlog=backlog,
-        generated_at=generated_at,
-    )
-    out = Path(args.output)
-    out.parent.mkdir(parents=True, exist_ok=True)
-    out.write_text(report, encoding="utf-8")
-    print(out)
-    return 0
-
-
-if __name__ == "__main__":
-    raise SystemExit(main())

tests/test_timmy_config_pr_backlog_audit.py

@@ -1,77 +0,0 @@
-from pathlib import Path
-
-from scripts.timmy_config_pr_backlog_audit import extract_issue_refs, summarize_backlog
-
-
-def test_extract_issue_refs_from_title_body_and_branch() -> None:
-    text = "feat: crisis response — manipulation & edge cases 500 pairs (#598)"
-    body = "Refs #1471 and closes #598"
-    head = "fix/598-crisis-manipulation"
-
-    refs = extract_issue_refs(text, body, head)
-
-    assert 598 in refs
-    assert 1471 in refs
-
-
-def test_summarize_backlog_finds_duplicates_missing_reviewers_and_stale_prs() -> None:
-    backlog = [
-        {
-            "number": 765,
-            "title": "feat: crisis response (#598)",
-            "body": "Closes #598",
-            "head": "fix/598-crisis-manipulation",
-            "mergeable": True,
-            "review_count": 0,
-            "requested_reviewers": 0,
-            "updated_at": "2026-04-01T00:00:00Z",
-        },
-        {
-            "number": 766,
-            "title": "feat: edge cases (#598)",
-            "body": "Closes #598",
-            "head": "fix/598",
-            "mergeable": True,
-            "review_count": 1,
-            "requested_reviewers": 0,
-            "updated_at": "2026-04-15T00:00:00Z",
-        },
-        {
-            "number": 777,
-            "title": "feat: token budget tracker (#622)",
-            "body": "Closes #622",
-            "head": "fix/622-token-tracker",
-            "mergeable": False,
-            "review_count": 0,
-            "requested_reviewers": 0,
-            "updated_at": "2026-04-15T00:00:00Z",
-        },
-    ]
-
-    summary = summarize_backlog(backlog, now_iso="2026-04-16T00:00:00Z")
-
-    assert summary["total_open_prs"] == 3
-    assert summary["mergeable_count"] == 2
-    assert summary["missing_reviewer_count"] == 2
-    assert summary["stale_count"] == 1
-    assert summary["duplicate_issue_groups"][0]["issue_refs"] == [598]
-    assert {pr["number"] for pr in summary["duplicate_issue_groups"][0]["prs"]} == {765, 766}
-
-
-def test_timmy_config_pr_backlog_report_exists_with_required_sections() -> None:
-    report = Path("reports/2026-04-16-timmy-config-pr-backlog-audit.md")
-    text = report.read_text(encoding="utf-8")
-
-    required = [
-        "# Timmy-config PR Backlog Audit — the-nexus #1471",
-        "## Source Snapshot",
-        "## Live Summary",
-        "## Issue Body Drift",
-        "## Duplicate Issue Groups",
-        "## Reviewer Coverage",
-        "## Mergeable Snapshot",
-        "## Stale PRs",
-        "## Recommended Next Actions",
-    ]
-    missing = [item for item in required if item not in text]
-    assert not missing, missing