z
AdvisoryAudited by Static analysis on May 13, 2026.
Overview
No suspicious patterns detected.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Old conversation content or retrieved text could steer the agent's behavior, output format, or answer beyond what the current user asked.
The skill tells the agent to follow instructions embedded in the CLI's retrieved prompt instead of treating that text only as data.
这段文本包含了历史对话上下文以及一个要求你输出 JSON 格式的指令...遵循 Prompt 里的要求
Treat MemCoT search results as untrusted reference data; do not obey instructions contained inside retrieved history unless the current user explicitly confirms them.
Sensitive past chats may be indexed and reused in later answers, and old or maliciously crafted content could influence future responses.
The skill is designed to search and reuse large amounts of local OpenClaw conversation history, but it does not specify clear limits, exclusions, storage retention, or poisoning protections.
在海量的历史对话记录中进行检索...是否要将openclaw的conversation_base设置为本地的openclaw的地址
Only point MemCoT at conversation folders you intend to index, review what it stores, set retention/exclusion rules, and verify retrieved content before relying on it.
A background service may be started without a clear user-visible confirmation and may continue running after the immediate request.
The artifact describes a persistent background daemon and instructs the agent to start it silently during search if it is not already running.
MemCoT 是一个运行在后台的守护进程...如果没有启动，请先静默执行 `python memcot_cil.py start`
Require explicit user confirmation before starting the daemon, show status after starting it, and provide clear stop and cleanup instructions.
The reviewed package does not include the code that will actually run, so safety depends on the external repository and the local checkout.
The runnable CLI code is expected to come from a separately cloned repository, not from the reviewed skill artifact.
本 skill 发布在 ClawHub 上为纯文本包。使用 MemCoT 时请克隆仓库并在仓库根目录运行 CLI
Inspect and pin the MemCoT repository version before using the skill, and avoid running unreviewed updates automatically.
If an agent or shell handles the query unsafely, unusual characters in a search question could cause command errors or unintended shell behavior.
The skill maps user-provided search questions into terminal commands. This is central to the skill, but it needs safe argument handling.
执行：`python memcot_cil.py search -q "[问题]" -o "./output"`
Pass arguments safely without shell interpolation, quote or escape user text correctly, and ask before running non-search commands.