5fd8346ef9
hook_blocks_export.py выгружает блокировки хуков в JSONL (cwd, ветка, команда, ответ хука) — сырьё для регрессионных тестов B2. Похожее на секреты скрывается. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
72 lines
3.7 KiB
Python
72 lines
3.7 KiB
Python
#!/usr/bin/env python3
|
|
"""Выгрузка блокировок хуков из транскриптов Claude Code в JSONL — сырьё для регрессионных тестов хуков (B2).
|
|
|
|
Запуск: python3 hook_blocks_export.py <каталог-транскриптов> [<каталог> ...] > blocks.jsonl
|
|
Одна строка — одна блокировка: сессия, время, cwd и ветка на момент вызова, инструмент, команда или путь файла,
|
|
первая строка ответа хука. Поле expected не заполняется: вердикт (block/allow) проставляется при разметке.
|
|
Похожие на секреты подстроки заменяются на <redacted>.
|
|
"""
|
|
import json, sys, glob, os, re
|
|
|
|
RE_SECRET = re.compile(
|
|
r'(?i)gh[pousr]_[A-Za-z0-9]{20,}|sk-[A-Za-z0-9_-]{20,}|\b[0-9a-f]{40}\b'
|
|
r'|(?P<kv>(?:token|password|passwd|secret|api[_-]?key)\s*[=:]\s*)\S+'
|
|
r'|(?P<flag>--(?:token|password)\s+)\S+|(?P<auth>Authorization:\s*\S+\s+)\S+'
|
|
)
|
|
|
|
|
|
def redact(s):
|
|
def sub(m):
|
|
keep = m.group("kv") or m.group("flag") or m.group("auth") or ""
|
|
return keep + "<redacted>"
|
|
return RE_SECRET.sub(sub, s)
|
|
|
|
|
|
def hook_of(txt):
|
|
if "shell write" in txt:
|
|
return "guard-main(Bash)"
|
|
if "you are on" in txt:
|
|
return "guard-main(Edit/Write)"
|
|
if "without a plan" in txt:
|
|
return "require-plan"
|
|
if "commit/push в main" in txt or "ветке main" in txt or "reference-old-arm" in txt or "docs/source" in txt:
|
|
return "workspace-guard"
|
|
return "other"
|
|
|
|
|
|
for d in sys.argv[1:]:
|
|
for f in sorted(glob.glob(os.path.join(os.path.expanduser(d), "*.jsonl"))):
|
|
sid = os.path.basename(f)[:8]
|
|
pending = {} # tool_use_id -> (name, input, cwd, branch, ts)
|
|
for line in open(f, encoding="utf-8", errors="replace"):
|
|
try:
|
|
j = json.loads(line)
|
|
except Exception:
|
|
continue
|
|
m = j.get("message") or {}
|
|
cont = m.get("content")
|
|
if j.get("type") == "assistant":
|
|
for b in cont or []:
|
|
if isinstance(b, dict) and b.get("type") == "tool_use":
|
|
pending[b.get("id")] = (b.get("name"), b.get("input") or {}, j.get("cwd", ""),
|
|
j.get("gitBranch", ""), j.get("timestamp", ""))
|
|
elif j.get("type") == "user" and isinstance(cont, list):
|
|
for b in cont:
|
|
if not (isinstance(b, dict) and b.get("type") == "tool_result"):
|
|
continue
|
|
txt = b.get("content", "")
|
|
if isinstance(txt, list):
|
|
txt = "\n".join(x.get("text", "") for x in txt if isinstance(x, dict))
|
|
if not ("BLOCKED" in txt and ("hook" in txt.lower() or "BLOCKED:" in txt[:400])):
|
|
continue
|
|
name, inp, cwd, branch, ts = pending.get(b.get("tool_use_id"), ("?", {}, "", "", ""))
|
|
if name not in ("Bash", "Edit", "Write", "MultiEdit"):
|
|
continue
|
|
msg = next((l for l in txt.splitlines() if "BLOCKED" in l), "")
|
|
print(json.dumps({
|
|
"session": sid, "ts": ts, "cwd": cwd, "branch": branch, "tool": name,
|
|
"command": redact(inp.get("command", "")) if name == "Bash" else None,
|
|
"file_path": inp.get("file_path") if name != "Bash" else None,
|
|
"hook": hook_of(txt), "message": redact(msg[:300]), "expected": None,
|
|
}, ensure_ascii=False))
|