Openclaw Starter
AdvisoryAudited by Static analysis on Apr 30, 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.
If the user installs all recommended skills without review, OpenClaw may gain broader automation capabilities than the user intended.
The skill recommends bulk installing automation and communication skills. This is disclosed and user-directed, but it can expand the agent's future ability to schedule tasks, handle webhooks, send email, or manage calendar/notification workflows.
clawhub install cron-manager reminder-skill webhook-handler email-sender calendar-manager notification-skill
Review each recommended skill and its permissions before installing, especially automation, email, webhook, SSH, Docker, or code-running tools.
A leaked or over-permissioned bot token could allow someone to read or post messages through the configured bot.
The Discord setup guide asks the user to create a bot token and grant message permissions. This is expected for channel integration, but those credentials and permissions are sensitive.
点击 Reset Token 获取 Bot Token ... 在权限中选择：Send Messages, Read Message History
Use dedicated bot accounts, grant only needed permissions, keep tokens private, and rotate any token that may have been exposed.
Following the optional dependency install may pull a newer dependency version than the author originally tested.
The optional Python dependency is specified with a version range rather than an exact pinned version. This is common, but it means future installs may resolve to different package versions.
requests>=2.28.0
If you need strict reproducibility, pin dependencies or install in a virtual environment before running the helper scripts.
Running the helper script will execute local status/version commands and print local environment information.
The helper script runs fixed local CLI commands to inspect OpenClaw, Node/npm, and installed skills. This is purpose-aligned environment checking and no shell injection or hidden execution is evidenced.
subprocess.run(['openclaw', 'gateway', 'status'], capture_output=True, text=True, timeout=10)
Run the helper scripts only from the installed skill directory you trust, and avoid sharing command output if it contains sensitive local details.
Information placed in USER.md, SOUL.md, or memory files may be reused by the assistant in later sessions.
The skill encourages creating persistent user/profile and memory files. This is aligned with OpenClaw onboarding, but persistent context can store private information and influence future assistant behavior.
USER.md - 关于你的信息 ... 记录你的名字、时区、偏好等 ... memory/YYYY-MM-DD.md - 每日记忆
Do not put secrets, tokens, or highly sensitive personal data in persistent profile or memory files, and review them periodically.