diff --git a/.config/nextest.toml b/.config/nextest.toml
new file mode 100644
index 0000000..20d6266
--- /dev/null
+++ b/.config/nextest.toml
@@ -0,0 +1,26 @@
+# cargo-nextest configuration. Shared by local runs and CI.
+#
+# Flakiness policy: no retries anywhere. A test that needs a retry is a bug
+# in the test; fix the test. `slow-timeout` turns a hang into a failure with
+# a name attached instead of a stalled job.
+
+[profile.default]
+slow-timeout = { period = "30s", terminate-after = 4 } # 2 min hard cap per test
+fail-fast = false
+
+[profile.ci]
+slow-timeout = { period = "30s", terminate-after = 6 } # 3 min on slower runners
+fail-fast = false
+failure-output = "immediate-final"
+status-level = "fail"
+final-status-level = "slow"
+
+[profile.ci.junit]
+path = "junit.xml"
+
+# Repeated-run profile for flake hunting: same policy, quieter output.
+[profile.stress]
+slow-timeout = { period = "30s", terminate-after = 4 }
+fail-fast = false
+status-level = "fail"
+final-status-level = "fail"
diff --git a/.github/scripts/coverage_summary.py b/.github/scripts/coverage_summary.py
new file mode 100755
index 0000000..72408d4
--- /dev/null
+++ b/.github/scripts/coverage_summary.py
@@ -0,0 +1,129 @@
+#!/usr/bin/env python3
+"""Render a Markdown coverage table from `cargo llvm-cov report --json`.
+
+Usage: coverage_summary.py HEAD.json [BASE.json]
+
+Rows are grouped per workspace crate (derived from the source path), with the
+line-coverage delta against BASE when given. Files under `target/` and
+`examples/`/`tests/` fixtures are still counted (they are part of the
+workspace) but examples are listed separately so library crates stay visible.
+"""
+import json
+import os
+import re
+import sys
+from collections import defaultdict
+
+CRATE_RE = re.compile(
+ r"(?:^|/)(crates/rpc/bidirectional/[^/]+|crates/[^/]+/[^/]+|examples/[^/]+(?:/[^/]+)?|tests/playwright/fixtures/[^/]+)/"
+)
+
+
+def load(path):
+ if not path or not os.path.exists(path):
+ return None
+ with open(path) as f:
+ data = json.load(f)
+ return data.get("data", [None])[0]
+
+
+def per_crate(report):
+ """-> {crate: (covered_lines, total_lines)}"""
+ out = defaultdict(lambda: [0, 0])
+ if not report:
+ return out
+ for f in report.get("files", []):
+ m = CRATE_RE.search(f["filename"])
+ crate = m.group(1) if m else "(other)"
+ lines = f["summary"]["lines"]
+ out[crate][0] += lines["covered"]
+ out[crate][1] += lines["count"]
+ return out
+
+
+def pct(cov, tot):
+ return 100.0 * cov / tot if tot else 0.0
+
+
+def fmt_delta(d):
+ if d is None:
+ return ""
+ if abs(d) < 0.005:
+ return "±0.00"
+ return f"{d:+.2f}"
+
+
+def main():
+ head = load(sys.argv[1])
+ base = load(sys.argv[2]) if len(sys.argv) > 2 else None
+ if head is None:
+ print("Coverage report unavailable.")
+ return
+
+ hc = per_crate(head)
+ bc = per_crate(base) if base else None
+
+ total_lines = head["totals"]["lines"]
+ total_pct = total_lines["percent"]
+ total_delta = None
+ if base:
+ total_delta = total_pct - base["totals"]["lines"]["percent"]
+
+ print("### Test coverage (lines)")
+ print()
+ headline = f"**Total: {total_pct:.2f}%**"
+ if total_delta is not None:
+ headline += f" ({fmt_delta(total_delta)} vs base)"
+ print(headline, f"— {total_lines['covered']}/{total_lines['count']} lines")
+ print()
+ cols = "| Crate | Lines | Coverage |" + (" Δ |" if base else "")
+ print(cols)
+ print("|---|---:|---:|" + ("---:|" if base else ""))
+
+ def rows(prefix):
+ for crate in sorted(k for k in hc if k.startswith(prefix)):
+ cov, tot = hc[crate]
+ p = pct(cov, tot)
+ line = f"| `{crate}` | {cov}/{tot} | {p:.2f}% |"
+ if base:
+ if crate in bc and bc[crate][1]:
+ d = p - pct(*bc[crate])
+ line += f" {fmt_delta(d)} |"
+ else:
+ line += " new |"
+ print(line)
+
+ rows("crates/")
+ if any(k.startswith(("examples/", "tests/")) for k in hc):
+ print("| **Examples and fixtures** | | |" + (" |" if base else ""))
+ rows("examples/")
+ rows("tests/")
+ if "(other)" in hc:
+ rows("(other)")
+
+ # Files whose coverage dropped the most, to make regressions actionable.
+ if base:
+ base_files = {f["filename"]: f["summary"]["lines"] for f in base.get("files", [])}
+ drops = []
+ for f in head.get("files", []):
+ b = base_files.get(f["filename"])
+ if not b or not b["count"]:
+ continue
+ d = f["summary"]["lines"]["percent"] - b["percent"]
+ if d <= -1.0:
+ drops.append((d, f["filename"], f["summary"]["lines"]["percent"]))
+ if drops:
+ print()
+ print("Files with coverage drops ≥ 1 point
")
+ print()
+ print("| File | Coverage | Δ |")
+ print("|---|---:|---:|")
+ for d, name, p in sorted(drops)[:25]:
+ short = name.split("/rust-api-stack/", 1)[-1]
+ print(f"| `{short}` | {p:.2f}% | {fmt_delta(d)} |")
+ print()
+ print("