|
| 1 | +#!/usr/bin/env python3 |
| 2 | +"""Sanitized Codex hook recorder for trading-runtime observability.""" |
| 3 | + |
| 4 | +from __future__ import annotations |
| 5 | + |
| 6 | +import argparse |
| 7 | +import datetime as dt |
| 8 | +import hashlib |
| 9 | +import json |
| 10 | +import os |
| 11 | +import re |
| 12 | +import sys |
| 13 | +from pathlib import Path |
| 14 | +from typing import Any |
| 15 | + |
| 16 | +MAX_TEXT = 4000 |
| 17 | +SECRET_PATTERNS = [ |
| 18 | + re.compile(r"(?i)(api[_-]?key|bearer|private[_-]?key|secret|token|seed|mnemonic)(['\"\s:=]+)[^\s'\",}]+"), |
| 19 | + re.compile(r"\b[1-9A-HJ-NP-Za-km-z]{80,}\b"), |
| 20 | +] |
| 21 | +SENSITIVE_KEYS = re.compile(r"(?i)(api[_-]?key|auth|bearer|mnemonic|private|secret|seed|token)") |
| 22 | +SIGNATURE_RE = re.compile(r"\b[1-9A-HJ-NP-Za-km-z]{64,88}\b") |
| 23 | + |
| 24 | + |
| 25 | +def now_utc() -> str: |
| 26 | + return dt.datetime.now(dt.UTC).isoformat().replace("+00:00", "Z") |
| 27 | + |
| 28 | + |
| 29 | +def read_input() -> dict[str, Any]: |
| 30 | + raw = sys.stdin.read() |
| 31 | + if not raw.strip(): |
| 32 | + return {} |
| 33 | + try: |
| 34 | + value = json.loads(raw) |
| 35 | + except json.JSONDecodeError: |
| 36 | + return {"raw_stdin": redact_text(raw)} |
| 37 | + return value if isinstance(value, dict) else {"value": value} |
| 38 | + |
| 39 | + |
| 40 | +def redact_text(value: str) -> str: |
| 41 | + value = value[:MAX_TEXT] |
| 42 | + for pattern in SECRET_PATTERNS: |
| 43 | + value = pattern.sub(redact_match, value) |
| 44 | + return value |
| 45 | + |
| 46 | + |
| 47 | +def redact_match(match: re.Match[str]) -> str: |
| 48 | + if match.lastindex and match.lastindex >= 2: |
| 49 | + return f"{match.group(1)}{match.group(2)}[REDACTED]" |
| 50 | + return "[REDACTED_LONG_BASE58]" |
| 51 | + |
| 52 | + |
| 53 | +def sanitize(value: Any) -> Any: |
| 54 | + if isinstance(value, dict): |
| 55 | + return { |
| 56 | + str(key): "[REDACTED]" if SENSITIVE_KEYS.search(str(key)) else sanitize(item) |
| 57 | + for key, item in value.items() |
| 58 | + } |
| 59 | + if isinstance(value, list): |
| 60 | + return [sanitize(item) for item in value[:50]] |
| 61 | + if isinstance(value, str): |
| 62 | + return redact_text(value) |
| 63 | + return value |
| 64 | + |
| 65 | + |
| 66 | +def event_summary(event: str, payload: dict[str, Any]) -> dict[str, Any]: |
| 67 | + summary: dict[str, Any] = {"event": event} |
| 68 | + for key in ("session_id", "turn_id", "cwd", "hook_event_name", "source"): |
| 69 | + if payload.get(key): |
| 70 | + summary[key] = payload[key] |
| 71 | + if prompt := payload.get("prompt"): |
| 72 | + summary["prompt_sha256"] = hashlib.sha256(str(prompt).encode()).hexdigest() |
| 73 | + summary["prompt_chars"] = len(str(prompt)) |
| 74 | + if tool_name := payload.get("tool_name"): |
| 75 | + summary["tool_name"] = tool_name |
| 76 | + if tool_input := payload.get("tool_input"): |
| 77 | + command = tool_input.get("command") if isinstance(tool_input, dict) else None |
| 78 | + if command: |
| 79 | + summary["command"] = redact_text(str(command)) |
| 80 | + tool_response = payload.get("tool_response") |
| 81 | + if isinstance(tool_response, dict): |
| 82 | + output = json.dumps(tool_response, sort_keys=True, default=str) |
| 83 | + signatures = sorted(set(SIGNATURE_RE.findall(output))) |
| 84 | + if signatures: |
| 85 | + summary["signatures"] = signatures[:10] |
| 86 | + summary["tool_response_chars"] = len(output) |
| 87 | + return summary |
| 88 | + |
| 89 | + |
| 90 | +def append_jsonl(path: Path, row: dict[str, Any]) -> None: |
| 91 | + path.parent.mkdir(parents=True, exist_ok=True) |
| 92 | + with path.open("a", encoding="utf-8") as handle: |
| 93 | + handle.write(json.dumps(row, sort_keys=True) + "\n") |
| 94 | + |
| 95 | + |
| 96 | +def append_markdown(path: Path, row: dict[str, Any]) -> None: |
| 97 | + path.parent.mkdir(parents=True, exist_ok=True) |
| 98 | + summary = row["summary"] |
| 99 | + parts = [f"- `{row['timestamp']}` `{row['event']}`"] |
| 100 | + if summary.get("tool_name"): |
| 101 | + parts.append(f"tool `{summary['tool_name']}`") |
| 102 | + if summary.get("prompt_chars") is not None: |
| 103 | + parts.append(f"prompt chars `{summary['prompt_chars']}`") |
| 104 | + if summary.get("signatures"): |
| 105 | + parts.append("signatures `" + "`, `".join(summary["signatures"]) + "`") |
| 106 | + with path.open("a", encoding="utf-8") as handle: |
| 107 | + handle.write(" ".join(parts) + "\n") |
| 108 | + |
| 109 | + |
| 110 | +def codex_json_summary(payload: dict[str, Any]) -> dict[str, Any]: |
| 111 | + summary: dict[str, Any] = {"event": "CodexExecJson"} |
| 112 | + if event_type := payload.get("type"): |
| 113 | + summary["codex_event_type"] = event_type |
| 114 | + item = payload.get("item") |
| 115 | + if isinstance(item, dict): |
| 116 | + summary["item_type"] = item.get("type") |
| 117 | + if text := item.get("text"): |
| 118 | + summary["text_sha256"] = hashlib.sha256(str(text).encode()).hexdigest() |
| 119 | + summary["text_chars"] = len(str(text)) |
| 120 | + if usage := payload.get("usage"): |
| 121 | + summary["usage"] = usage |
| 122 | + return summary |
| 123 | + |
| 124 | + |
| 125 | +def stream_codex_jsonl(root: Path) -> int: |
| 126 | + for line in sys.stdin: |
| 127 | + if not line.strip(): |
| 128 | + continue |
| 129 | + try: |
| 130 | + raw_payload = json.loads(line) |
| 131 | + except json.JSONDecodeError: |
| 132 | + raw_payload = {"raw_line": line} |
| 133 | + payload = sanitize(raw_payload) |
| 134 | + timestamp = now_utc() |
| 135 | + row = { |
| 136 | + "timestamp": timestamp, |
| 137 | + "event": "CodexExecJson", |
| 138 | + "summary": codex_json_summary(raw_payload), |
| 139 | + "payload": payload, |
| 140 | + } |
| 141 | + append_jsonl(root / "codex-exec-events.jsonl", row) |
| 142 | + append_markdown(root / f"{timestamp[:10]}.md", row) |
| 143 | + print(json.dumps(payload, sort_keys=True), flush=True) |
| 144 | + return 0 |
| 145 | + |
| 146 | + |
| 147 | +def main() -> int: |
| 148 | + parser = argparse.ArgumentParser() |
| 149 | + parser.add_argument("--event", required=True) |
| 150 | + parser.add_argument("--stream-jsonl", action="store_true") |
| 151 | + args = parser.parse_args() |
| 152 | + |
| 153 | + root = Path(os.environ.get("CODEX_OBSERVABILITY_DIR", ".agent-observability")) |
| 154 | + if args.stream_jsonl: |
| 155 | + return stream_codex_jsonl(root) |
| 156 | + |
| 157 | + raw_payload = read_input() |
| 158 | + payload = sanitize(raw_payload) |
| 159 | + timestamp = now_utc() |
| 160 | + row = { |
| 161 | + "timestamp": timestamp, |
| 162 | + "event": args.event, |
| 163 | + "summary": event_summary(args.event, raw_payload), |
| 164 | + "payload": payload, |
| 165 | + } |
| 166 | + append_jsonl(root / "events.jsonl", row) |
| 167 | + append_markdown(root / f"{timestamp[:10]}.md", row) |
| 168 | + return 0 |
| 169 | + |
| 170 | + |
| 171 | +if __name__ == "__main__": |
| 172 | + raise SystemExit(main()) |
0 commit comments