Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Session Archiver

Session auto-archiver. Automatically archives completed sessions by extracting meaningful user+assistant messages from .reset transcript files and appending...

MIT-0 · Free to use, modify, and redistribute. No attribution required.
0 · 90 · 3 current installs · 3 all-time installs
bysniper-one@aqbjqtd
MIT-0
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description (session auto-archiver) match the scripts and SKILL.md: both scripts scan the sessions directory, parse .reset jsonl transcripts, strip metadata, deduplicate, append user+assistant text to memory/YYYY-MM-DD.md, and optionally back up .reset files to WebDAV.
Instruction Scope
Runtime instructions are explicit: spawn a subagent on /new to run the provided node script. The scripts read session transcript files and memory files (expected for archiving), remove metadata, generate insight candidates, and optionally perform network calls for WebDAV. They do not attempt to read unrelated system config or external services beyond optional WebDAV.
Install Mechanism
No install spec or external downloads; this is instruction+bundled script only. No remote code is fetched or extracted at install time.
Credentials
The registry lists no required env vars, but SKILL.md and scripts optionally use WEBDAV_ENABLED, WEBDAV_INTERNAL, WEBDAV_EXTERNAL, and WEBDAV_AUTH for backups. These are proportional to the optional WebDAV feature, but WEBDAV_AUTH is sensitive (user:app_password) and should be supplied only if you trust the target WebDAV endpoint.
Persistence & Privilege
always:false and no attempts to modify other skills or system-wide settings. The script writes only into the workspace/memory area and the marker file it manages; it also deletes old memory files per its retention policy — that behavior is documented and consistent with its purpose.
Assessment
This skill is coherent with its description: it will read your agent .reset transcript files and append extracted user+assistant text to daily memory files, and (only if you enable it) upload raw .reset files to a WebDAV endpoint. Before installing: review the scripts (they're included) so you are comfortable with the exact paths and retention policy; do not set WEBDAV_AUTH unless you want backups sent to the configured WebDAV server; restrict access to the memory and sessions directories if they contain sensitive data; and if you are concerned about autonomous background runs, keep the skill user-invocable only (don't change platform-level auto-invoke settings).
scripts/session-auto-archive.js:53
Shell command execution detected (child_process).
Patterns worth reviewing
These patterns may indicate risky behavior. Check the VirusTotal and OpenClaw results above for context-aware analysis before installing.

Like a lobster shell, security has layers — review code before you run it.

Current versionv1.5.1
Download zip
latestvk97269w4ppr4agrhes1fct0m1183twz1

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

SKILL.md

Session Archiver

一句话

/new 时自动归档对话历史到每日 memory,保留 user + assistant 消息,可选 WebDAV 备份。

核心要点(4条)

  1. 触发 — Session Startup 第 5 步 spawn subagent 调用(/new 时)
  2. 去重 — marker + memory 内容双保险,避免重复归档
  3. 归档内容 — user + assistant 消息,彻底剥离元数据
  4. 自动清理 — 保留最近 30 天日志,超出自动删除
  5. 洞察提炼 — 自动提取值得沉淀的洞察候选,供下次会话审查

快速集成

安装后在 AGENTS.md 的 Session Startup 章节添加:

5. **自动沉淀(上一次 session)** — 后台 spawn subagent 运行 skill `session-archiver`,完全不阻塞主窗口
6. **洞察候选审查** — 检查 `memory/.insights-candidates-YYYY-MM-DD.md`,快速过一遍上次 session 的洞察候选,决定是否吸收进 self-improving 文件

Session Startup 第 5 步的 spawn 触发归档,第 6 步由主 agent 读取洞察候选文件。

脚本位置

skills/session-archiver/scripts/session-auto-archive.js

执行方式

/new 时自动触发(主流程)

sessions_spawn runtime="subagent" background=true
任务:Run session auto-archive script: node /home/node/.openclaw/workspace/skills/session-archiver/scripts/session-auto-archive.js

无需 cron,/new 即触发。

工作流程

  1. 检测 WebDAV 连通性(内部 → 外部fallback)
  2. 扫描 ~/.openclaw/agents/main/sessions/ 找最新 .reset.* 文件
  3. 去重 A:检查 memory/.last_archived_session marker
  4. 去重 B:检查 memory 文件是否已有该 sessionId
  5. 解析 JSONL,提取 user + assistant 消息
  6. 彻底剥离 Conversation info 元数据
  7. 短时内容相似去重(hash 前 80 字符)
  8. 追加到 memory/YYYY-MM-DD.md
  9. WebDAV 可达时上传 .reset 文件到配置的备份目录
  10. 更新 marker
  11. 洞察提炼 — 调用 extract-insights.js,结果写入 memory/.insights-candidates-YYYY-MM-DD.md
  12. 自动清理 — 删除 30 天前的 memory 文件

去重机制

检查点文件说明
markermemory/.last_archived_session记录上次处理的 sessionId
memory 内容memory/YYYY-MM-DD.md搜索 来源 session: <id>

WebDAV 配置(可选)

通过环境变量配置,不会上传到 ClawHub:

  • WEBDAV_ENABLED — 设为 true 启用 WebDAV 备份(默认 false,需手动开启)
  • WEBDAV_INTERNAL — 内部地址(容器间)
  • WEBDAV_EXTERNAL — 外部地址(公网)
  • WEBDAV_AUTH — 认证信息(格式:user:app_password

WEBDAV_ENABLED=false 时完全跳过 WebDAV 相关检测和上传。

输出格式

## Session 自动沉淀

### 来源 session(.reset 文件归档)

- **sessionId**: xxx
- **reset文件**: xxx.jsonl.reset.TIMESTAMP

### 对话内容(共 N 条)

👤 用户发送的实际内容...
🤖 Agent 的回答(有价值的)...
---
归档时间: ISO timestamp
来源 session: sessionId

洞察候选文件

extract-insights.js 分析归档内容,识别 6 类值得提炼的信息:

类型说明
🤕 纠正用户纠正了 AI 的行为
🔄 自我修正AI 承认错误或改进
📌 决策形成了明确结论或建议
💡 偏好用户表达了明确偏好
⚙️ 工作流变化任务分工、工具选择发生变化
📝 教训需要记住避免重蹈的规则

输出到 memory/.insights-candidates-YYYY-MM-DD.md,主 agent 在下次 Session Startup 时审查,决定是否吸收进 self-improving 文件。

注意事项

  • 仅沉淀 text 内容,图片/文件不内联存储
  • 长期重要内容需人工迁移至 MEMORY.md
  • 主窗口永远不直接运行脚本,只 spawn subagent
  • WebDAV 不可达时静默跳过,不影响归档流程
  • 洞察提炼为辅助功能,找不到候选时静默退出

Files

3 total
Select a file
Select a file to preview.

Comments

Loading comments…