fix(chains): drop node URLs that embed Nova's provider keys (#72)

* fix(chains): drop node URLs that embed Nova's Infura keys

Nova's configs carry Nova's own Infura project keys inside node URLs, and
every sync copied them into ours: Ethereum had four keyed Infura nodes and
Base one, across 43 published files. Our Android build already fills
`{INFURA_API_KEY}` from our own secret; the keyed URLs used another
organisation's quota and would break for every user when Nova rotates them.

- sync_from_nova.py strips nodes whose URL embeds an Infura key, and drops a
  chain left with no node (only Ethereum - Goerli, a retired testnet, in the
  dev files); merge-chains.py applies the same filter
- the published chain files are rewritten with that filter; nothing else in
  them changes (each file round-trips byte for byte through json.dump)
- check_pezkuwi_integrity.py refuses a keyed Infura URL under chains/

* fix(chains): strip every keyed provider URL, and make the gate fail closed

The first pass matched only `infura.io/.../v3/<lowercase hex>`, so it missed
`https://...infura.io/v3/<key>`, upper-case keys and hosts, and Nova's
OnFinality `?apikey=` URLs (Moonriver, and the Hades testnet in v20/v21
preConfigured). preConfigured was copied from Nova without the filter, and the
integrity check passed if grep itself failed.

- one predicate, is_keyed(), in sync_from_nova.py: Infura /v3/ ids and
  apikey/api_key/key/token query values, case-insensitive; {PLACEHOLDER}
  values are allowed
- preConfigured details are filtered after the copy; a chain left with no
  node loses its details file and its index entry (Hades, v20/v21 only)
- rewritten files keep their own indent, so the diff is the data change only
- check 5 imports is_keyed() and checks every string in every JSON file
  under chains/; an unreadable file or an empty tree is an error
- check 3 reports a grep error instead of passing on it

scripts/check_pezkuwi_integrity.py

@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
-"""Refuse the four regressions this repo has already had.
+"""Refuse the five regressions this repo has already had.
 
 Each check below exists because the thing it forbids happened, was fixed, and in two
 cases came back. They are cheap; the failures they catch are not.
@@ -88,6 +88,8 @@ grep = subprocess.run(
     capture_output=True,
     text=True,
 )
+if grep.returncode not in (0, 1):
+    errors.append(f"grep for working-branch asset URLs failed: {grep.stderr.strip()}")
 if grep.stdout.strip():
     files = grep.stdout.strip().splitlines()
     errors.append(
@@ -107,6 +109,43 @@ for fixture in ("pezkuwi_assets_for_testBalance.json",):
             f"tests/{fixture} has drifted from pezkuwi-overlay/tests/{fixture}"
         )
 
+# ── 5. No third-party API key in a URL ───────────────────────────────────────
+#
+# Nova's configs shipped Nova's own provider keys (Infura project ids, OnFinality
+# apikeys) inside node URLs, and every sync copied them into ours. Our builds fill
+# {INFURA_API_KEY}-style placeholders from our own secrets; a keyed URL is someone
+# else's quota and dies when they rotate it. The predicate is the one the sync
+# uses, and every string in every JSON file is checked, so a file that does not
+# parse is a failure, not a pass.
+sys.path.insert(0, str(ROOT / "scripts"))
+from sync_from_nova import is_keyed  # noqa: E402
+
+keyed_files = []
+json_files = sorted(p for p in (ROOT / "chains").rglob("*.json"))
+if not json_files:
+    errors.append("chains/ has no JSON files; the key check would have checked nothing")
+for path in json_files:
+    try:
+        doc = json.loads(path.read_text(encoding="utf-8"))
+    except (OSError, ValueError) as exc:
+        errors.append(f"{path.relative_to(ROOT)} could not be read for the key check: {exc}")
+        continue
+    stack = [doc]
+    while stack:
+        node = stack.pop()
+        if isinstance(node, dict):
+            stack.extend(node.values())
+        elif isinstance(node, list):
+            stack.extend(node)
+        elif isinstance(node, str) and is_keyed(node):
+            keyed_files.append(path)
+            break
+if keyed_files:
+    errors.append(
+        f"{len(keyed_files)} file(s) under chains/ embed a provider key in a URL; "
+        f"use a {{PLACEHOLDER}}, e.g. {keyed_files[0].relative_to(ROOT)}"
+    )
+
 if errors:
     for e in errors:
         print(f"::error::{e}")
@@ -114,4 +153,5 @@ if errors:
     sys.exit(1)
 
 print("Pezkuwi integrity: on-chain assets only, sufficiency declared, "
-      "assets served from master, fixture in step with its source.")
+      "assets served from master, fixture in step with its source, "
+      "no third-party key in a node URL.")

scripts/merge-chains.py

@@ -15,8 +15,12 @@ Options:
 
 import json
 import argparse
+import sys
 from pathlib import Path
 
+sys.path.insert(0, str(Path(__file__).parent))
+from sync_from_nova import strip_keyed_endpoints  # noqa: E402
+
 # Base paths
 SCRIPT_DIR = Path(__file__).parent
 ROOT_DIR = SCRIPT_DIR.parent
@@ -151,6 +155,9 @@ def merge_chains(nova_chains: list, pezkuwi_chains: list, filter_broken: bool =
         nova_filtered.append(chain)
         stats['nova_included'] += 1
 
+    nova_filtered = strip_keyed_endpoints(nova_filtered)
+    stats['nova_included'] = len(nova_filtered)
+
     # Pezkuwi first, then Nova
     merged = pezkuwi_chains + nova_filtered
     stats['total'] = len(merged)

scripts/sync_from_nova.py

@@ -19,6 +19,7 @@ Rules:
 """
 
 import json
+import re
 import shutil
 from pathlib import Path
 
@@ -51,6 +52,67 @@ def load_blacklist() -> set:
     return set()
 
 
+# Nova's configs carry Nova's own provider keys written into node URLs (Infura
+# project ids, OnFinality apikeys). They are not ours to use: our builds fill
+# placeholders like `{INFURA_API_KEY}` from our own secrets, and a keyed URL
+# breaks for every user the day Nova rotates it. check_pezkuwi_integrity.py
+# imports this same predicate, so the filter and the gate cannot disagree.
+KEYED_ENDPOINT = re.compile(
+    r"infura\.io/(?:[^?#\s]*/)?v3/[0-9a-z]{32}"
+    r"|[?&](?:api[_-]?key|apikey|access[_-]?token|token|key)=(?!\{)[^&#\s]+",
+    re.IGNORECASE,
+)
+
+
+def is_keyed(url: str) -> bool:
+    """True if the URL embeds a literal credential rather than a {PLACEHOLDER}."""
+    return bool(KEYED_ENDPOINT.search(url))
+
+
+def strip_chain(chain: dict) -> dict | None:
+    """The chain without keyed nodes; None if no node is left."""
+    nodes = chain.get('nodes')
+    if not isinstance(nodes, list):
+        return chain
+    kept = [n for n in nodes if not (isinstance(n, dict) and is_keyed(str(n.get('url', ''))))]
+    if not kept:
+        return None
+    return chain if len(kept) == len(nodes) else {**chain, 'nodes': kept}
+
+
+def strip_keyed_endpoints(chains: list) -> list:
+    """Drop nodes whose URL embeds a third-party key, and any chain left with no node."""
+    return [c for c in (strip_chain(c) for c in chains) if c is not None]
+
+
+def rewrite_in_place(path: Path, data):
+    """Rewrite a JSON file keeping its own indent and trailing newline, so a
+    filtered copy of Nova's file differs from it only where the data does."""
+    raw = path.read_text(encoding='utf-8')
+    second = raw.split('\n', 2)[1] if raw.count('\n') >= 1 else ''
+    indent = (len(second) - len(second.lstrip(' '))) or 2
+    out = json.dumps(data, indent=indent, ensure_ascii=False)
+    path.write_text(out + ('\n' if raw.endswith('\n') else ''), encoding='utf-8')
+
+
+def strip_preconfigured(preconfig: Path):
+    """Apply strip_chain to every details file; a chain left with no node is
+    removed together with its entry in the matching index."""
+    for details, index in (("details", "chains.json"), ("detailsDev", "chains_dev.json")):
+        dropped = set()
+        for f in sorted((preconfig / details).glob("*.json")):
+            chain = load_json(f)
+            stripped = strip_chain(chain)
+            if stripped is None:
+                dropped.add(chain.get('chainId'))
+                f.unlink()
+            elif stripped is not chain:
+                rewrite_in_place(f, stripped)
+        index_file = preconfig / index
+        if dropped and index_file.exists():
+            rewrite_in_place(index_file, [c for c in load_json(index_file) if c.get('chainId') not in dropped])
+
+
 def merge_chains(nova_chains: list, pezkuwi_chains: list, blocked_ids: set) -> list:
     """Pezkuwi chains first, then Nova chains. Blocked chains excluded."""
     pezkuwi_ids = {c['chainId'] for c in pezkuwi_chains}
@@ -58,7 +120,7 @@ def merge_chains(nova_chains: list, pezkuwi_chains: list, blocked_ids: set) -> l
         c for c in nova_chains
         if c['chainId'] not in pezkuwi_ids and c['chainId'] not in blocked_ids
     ]
-    return pezkuwi_chains + nova_filtered
+    return pezkuwi_chains + strip_keyed_endpoints(nova_filtered)
 
 
 def merge_xcm(nova_xcm: dict, pezkuwi_xcm: dict) -> dict:
@@ -143,6 +205,7 @@ def sync_chains():
             if output_preconfig.exists():
                 shutil.rmtree(output_preconfig)
             shutil.copytree(nova_preconfig, output_preconfig)
+            strip_preconfigured(output_preconfig)
 
 
 def sync_xcm():