paper-parallel-audit — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited paper-parallel-audit (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
156 篇引用串行核查 ≈ 5 小时 + 中间挂了从头来。 3 个 Agent 并行 + 每个 50 篇 + JSON 落盘 ≈ 1 小时 + 任意一个挂了只重跑那一个。
批量任务的瓶颈不是 AI 的速度,是"挂了从头来"的恐惧。 并行 + 落盘 = 把恐惧拆成可控的小块。
满足全部 → 触发:
paper-pilot-before-batch 跑过样本,逻辑确认无误任意一条不满足 → 不并行,老老实实串行 + 落盘(仍然要落盘)。
total = 156 条
shard_size = 50(按 Agent 数倒推)
shards = [0:50, 50:100, 100:156]每个 Agent 处理自己的分片,输出:
{
"shard_id": "0-50",
"total": 50,
"results": [
{"item_id": 1, "status": "pass", "issue": null},
{"item_id": 2, "status": "fail", "issue": "作者名拼写错误", "suggestion": "..."}
],
"completed_at": "<ISO 8601 timestamp>"
}# 等所有 Agent 返回后
python merge_shards.py shard-*.json > audit_report.json检测到 ≥ 30 条目的同质批量任务
│
▼
确认已经跑过 paper-pilot-before-batch
│
▼
分片:N 个 Agent,每个 ≤ 50 条
│
▼
告诉用户:「派 N 个 Agent 并行,每个负责 [区间],
各自落盘 JSON。预计 [时间],挂了只重跑挂掉的分片。」
│
▼
用 Task 工具派 Agent(subagent_type 选 general-purpose)
单条消息里多个 Task 调用 = 真正并行
│
▼
等所有 Agent 返回 → 汇总 JSON → 给用户报告你是引用核查 Agent,负责本批 [分片区间]。
任务:核查每条引用的 [核查标准列表]。
输入:[引用列表 / 文件路径]
输出:JSON 文件,路径 `shard-<分片 ID>.json`,格式:
{
"shard_id": "...",
"total": N,
"results": [
{"item_id": ..., "status": "pass|fail", "issue": "...", "suggestion": "..."}
],
"completed_at": "..."
}
完成后只返回一句话:「分片 X 完成,pass M 条,fail K 条,已落盘 shard-X.json」。
不要把详细结果写在回复里——一律只在 JSON 里。用户:「把这 156 条引用核查一遍。」
错误做法:在主会话里串行跑——
正确做法:分 3 片 → 派 3 个 Agent → 各自 JSON 落盘 → 主进程汇总 → 用户拿到完整报告。
| 念头 | 现实 |
|---|---|
| "派 Agent 太重,自己串行更直接" | 串行第 80 条挂掉时你会回来读这个表 |
| "我能边跑边把结果记在脑子里" | 你不能。这是惨案 9 的根因 |
| "JSON 太麻烦,纯文本 markdown 就行" | 汇总时要 dedupe / 排序 / filter,markdown 干不了 |
| "派 1 个 Agent 跑 156 条也是并行" | 不是,那只是把串行搬到了子进程里,还是从头来 |
| "落盘的 JSON 占空间" | 一个 shard JSON 几 KB,挂了能省小时级时间 |
paper-pilot-before-batchpaper-protect-terminologypaper-verify-before-handoff 自查报告《Claude Code 科研手记》§4.2「批量文献核查」、§10.2「引用核查的并行实战」
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.