sno-ai/sno-station-skills / skills/J-peer-review-and-audit/peer-review/skill
peer-review
VerifiedAdversarial review and fix loop for code, specs, plans, and migration manifests.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 32.7 KB |
| agents/openai.yaml | config | 216 B |
| assets/adversarial-plan-review.prompt.md | markdown | 11.0 KB |
| assets/adversarial-review.prompt.md | markdown | 12.0 KB |
| assets/adversarial-test-review.prompt.md | markdown | 3.5 KB |
| references/code-layers.md | markdown | 5.4 KB |
| references/doc-layers.md | markdown | 5.5 KB |
| references/python-advanced-patterns.md | markdown | 7.8 KB |
| references/python-standards.md | markdown | 8.3 KB |
| references/ts-standards.md | markdown | 7.1 KB |
| scripts/lib/sno-model.sh | script | 1.5 KB |
| scripts/review-backfill.shexecutable | script | 4.7 KB |
| scripts/review-findings.shexecutable | script | 11.7 KB |
| scripts/review-receipt.shexecutable | script | 5.3 KB |
| scripts/run-adversarial-review.shexecutable | script | 57.6 KB |
| scripts/selftest/cursor-process-group.t | other | 2.0 KB |
| scripts/selftest/run-cross-vendor.sh | script | 14.5 KB |
| scripts/selftest/run-model-settings.shexecutable | script | 5.1 KB |
| scripts/selftest/run-selftest.shexecutable | script | 39.2 KB |
scripts/review-backfill.sh
View on GitHub#!/usr/bin/env bash
# review-backfill.sh — recover past Codex-reviewer reports from Codex's own session logs.
#
# review-backfill.sh [--since <YYYY-MM-DD>] [--dry-run]
#
# If report files were lost, the findings can still be recovered — every Codex
# review's final message is in ~/.codex/sessions, which is where this reads them
# from. Recovery is a scan, not a reconstruction.
#
# What it does: finds every rollout whose prompt is one of the adversarial-review
# templates, writes its final report into the archive under the date it actually
# ran, then hands the lot to review-findings.sh in one call. Re-running is safe —
# findings are keyed by file plus title, so a second pass updates rows instead of
# duplicating them, and an already-written report file is not rewritten.
set -Eeuo pipefail
SESSIONS="${CODEX_SESSIONS_DIR:-$HOME/.codex/sessions}"
ARCHIVE="${REVIEW_ARCHIVE_DIR:-$HOME/.local/state/codex-reviews}"
HERE="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd)"
since=""
dry=0
while [ $# -gt 0 ]; do
case "$1" in
--since) [ $# -ge 2 ] || { echo "review-backfill: --since needs a date" >&2; exit 2; }
since="$2"; shift 2 ;;
--dry-run) dry=1; shift ;;
-h|--help) sed -n '2,/^set -Eeuo pipefail$/p' "$0" | sed '$d;s/^# \{0,1\}//'; exit 0 ;;
*) echo "review-backfill: unknown arg $1" >&2; exit 2 ;;
esac
done
[ -d "$SESSIONS" ] || { echo "review-backfill: no session logs at $SESSIONS" >&2; exit 1; }
[ -n "$since" ] || since="$(date -d '30 days ago' +%F 2>/dev/null || date -v-30d +%F)"
echo "[review-backfill] scanning $SESSIONS for reviews since $since" >&2
manifest="$(mktemp)"
trap 'rm -f -- "$manifest"' EXIT
python3 - "$SESSIONS" "$ARCHIVE" "$since" "$dry" "$manifest" <<'PY'
import json, os, sys, glob, time
sessions, archive, since, dry, manifest = sys.argv[1:6]
dry = dry == "1"
paths = sorted(glob.glob(os.path.join(sessions, "*", "*", "*", "*.jsonl")))
print(f"[review-backfill] {len(paths)} session file(s) to scan", file=sys.stderr)
t0 = time.time()
written = skipped = scanned = 0
out_paths = []
for i, p in enumerate(paths):
if i and i % 1000 == 0:
el = time.time() - t0
print(f"[review-backfill] {i}/{len(paths)} {el:.0f}s {written} recovered",
file=sys.stderr, flush=True)
base = os.path.basename(p)
# rollout-<timestamp>-<uuid>.jsonl — cheap date filter before reading
day = base[8:18] if base.startswith("rollout-") else ""
if day and day < since:
continue
scanned += 1
is_review = False
last = None
ts = day
try:
with open(p, encoding="utf-8", errors="replace") as fh:
for line in fh:
if '"user_message"' in line and not is_review:
try:
msg = json.loads(line)["payload"].get("message", "")
except Exception:
continue
head = msg[:400].lower()
if "adversarial" in head and "review" in head:
is_review = True
elif '"agent_message"' in line:
try:
rec = json.loads(line)
last = rec["payload"].get("message", "")
ts = rec.get("timestamp", ts)[:10] or ts
except Exception:
pass
except OSError:
continue
if not is_review or not last:
continue
sid = base.rsplit("-", 5)[-1].replace(".jsonl", "") if "-" in base else base
dest_dir = os.path.join(archive, ts or "unknown-date")
dest = os.path.join(dest_dir, f"backfill-{sid}.md")
out_paths.append(dest)
if os.path.exists(dest):
skipped += 1
continue
if dry:
written += 1
continue
os.makedirs(dest_dir, exist_ok=True)
with open(dest, "w", encoding="utf-8") as f:
f.write(last)
written += 1
with open(manifest, "w") as f:
for p_ in out_paths:
f.write(p_ + "\n")
print(f"[review-backfill] scanned {scanned} session(s) in window: "
f"{written} report(s) {'would be ' if dry else ''}recovered, {skipped} already present",
file=sys.stderr)
PY
if [ "$dry" = 1 ]; then
echo "[review-backfill] dry run — nothing written, no findings recorded" >&2
exit 0
fi
count="$(wc -l < "$manifest" | tr -d ' ')"
if [ "$count" = 0 ]; then
echo "[review-backfill] no reports in the window; nothing to record" >&2
exit 0
fi
# One call, not one per report: the ledger is rewritten whole on each save.
# xargs -a and -d are GNU-only; NUL-separated input reads the same paths on GNU and BSD (macOS) xargs.
tr '\n' '\0' < "$manifest" | xargs -0 bash "$HERE/review-findings.sh" record --kind recovered
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |