Skill flagged — suspicious patterns detected

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

Context Continuity

v1.1.0

上下文续接与记忆管理。当用户说"新建对话"、"继续"、"接着上次"等时, 自动从 MEMORY.md 和 memory/ 文件中提取当前任务/项目上下文, 让新会话能够了解之前的进展。用于:(1) 用户要开始新对话但想保留上下文 (2) 用户说"继续上次的工作" (3) 用户想要回顾当前进行中的任务 特点:自动记...

0· 310·0 current·0 all-time
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The name/description (context continuity and memory management) match the instructions: the skill writes daily memory files and reads recent memory to resume context. This capability is coherent with the stated purpose. However, the SKILL.md expects a local memory directory and a workspace path (~/.openclaw/workspace/) even though the skill metadata declared no required config paths — a mismatch that should be clarified.
!
Instruction Scope
The runtime instructions directly tell the agent to append every conversation to memory/YYYY-MM-DD.md and to read the last 7 days of memory via shell commands (echo, cat, date). They mandate 'automatic recording of all conversations without user confirmation', which is a material privacy/scope decision. The instructions reference specific file paths (memory/, MEMORY.md, ~/.openclaw/workspace/) and perform file I/O; those file accesses are not declared in the skill metadata and could include sensitive content.
Install Mechanism
No install spec and no code files are present (instruction-only). This minimizes supply-chain/install risk because nothing is downloaded or written during install.
!
Credentials
The skill requests no environment variables or credentials, which is appropriate, but it nevertheless performs local file writes/reads in the user's workspace and memory directories. Because these paths were not declared in requires.config, the skill's access to local files is not proportionately disclosed. Automatically writing full conversation content to disk without opt-in is a privacy risk.
Persistence & Privilege
always is false and model invocation is allowed (defaults). There is no indication the skill tries to modify other skills or system-wide settings. The main persistence concern is the skill's own habit of creating/writing long-term memory files and recommending archiving/deletion policies; that behavior is powerful but consistent with its purpose.
What to consider before installing
This skill will append all conversations to local files (memory/YYYY-MM-DD.md and MEMORY.md under your workspace) and will read recent memory automatically when you ask it to 'continue'. Before installing, consider: (1) Do you consent to automatic, unprompted logging of every conversation? (2) Where will memory/ and ~/.openclaw/workspace/ be created and who can read those files? (3) Do you need an opt-in toggle, redaction, or encryption for stored memories? (4) Ask the publisher to declare config paths and to change 'auto-write without confirmation' to an opt-in setting or at least provide a clear on/off command. If you proceed, restrict filesystem access, secure or encrypt the memory directory, and test deletion/archival behavior so private data isn't retained unintentionally.

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

latestvk971sm55hxextnzs76ncrcba0s838sqp
310downloads
0stars
2versions
Updated 3h ago
v1.1.0
MIT-0

Context Continuity Skill

核心功能:自动记忆

无需用户确认,自动记录所有对话内容到日常记忆文件。

记录规则

  • 每个会话结束时,自动写入 memory/YYYY-MM-DD.md
  • 记录内容:任务、项目、决策、待办事项
  • 保持摘要简洁,但内容要完整

触发条件(续接上下文)

当用户说以下话时,自动提取记忆:

  • "新建对话"
  • "继续"
  • "接着上次"
  • "上次我们做什么了"
  • "继续之前的工作"
  • "回顾上下文"

执行步骤

1. 自动记录(每次对话)

当对话结束时:

# 获取当前日期
date +%Y-%m-%d

# 追加到 memory/YYYY-MM-DD.md
echo "- $(date '+%H:%M') | 对话内容摘要" >> memory/$(date +%Y-%m-%d).md

记录格式:

# 2026-03-19

## 21:05
- 用户说:继续之前的工作
- 提取上下文:用户正在开发购物平台

## 21:09
- 用户说:自动记忆更好
- 更新:改为自动记忆模式

2. 续接上下文(触发时)

当用户要求"继续"时:

# 读取最近7天的记忆
cat memory/$(date -d '6 days ago' +%Y-%m-%d).md
cat memory/$(date -d '5 days ago' +%Y-%m-%d).md
# ... 一直到今天

3. 提取关键信息

从记忆中找出:

  • 当前进行的任务/项目
  • 未完成的事项
  • 重要决策或进展
  • 用户偏好或设置

4. 生成上下文摘要

用 3-5 句话总结当前状态,例如:

"用户正在开发一个购物平台。今天我们完善了 context-continuity skill,改为自动记忆模式。之前的任务是..."

5. 呈现给用户

将摘要呈现后,继续对话。

文件位置

  • 长期记忆:MEMORY.md(重要/永久信息)
  • 日常记忆:memory/YYYY-MM-DD.md(每日对话记录)
  • 工作空间:~/.openclaw/workspace/

注意事项

  • 每次对话结束自动写入记忆
  • 保持摘要简洁,不超过 200 字
  • 如果没有记忆,告知用户"这是新会话,没有历史上下文"
  • 尊重用户隐私,不过滤内容(用户需要可自行删除)
  • 定期清理:超过30天的记忆可以归档或删除

Comments

Loading comments...