OpenClaw Minecraft

Security checks across malware telemetry and agentic risk

Overview

The skill is clearly for Minecraft bot control, but it asks to replace the workspace cron prompt and run an unbounded 30-second autonomous action loop through an external controller.

Install only if you intentionally want persistent Minecraft bot autonomy. Before enabling it, back up any existing CRON_PROMPT.md, verify the external controller endpoint, use a private or test Minecraft server, scope and protect the controller token, disable open registration unless needed, and add clear stop conditions for the cron job.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Risk analysis

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.

#
ASI01: Agent Goal Hijack
Medium
What this means

Existing cron instructions or other scheduled workflows could be clobbered, causing the agent to prioritize Minecraft automation every cron cycle.

Why it was flagged

This directs replacement of a global workspace cron instruction file even if existing automation is present, which can redirect scheduled agent behavior toward this skill.

Skill content
Overwrite the workspace root `CRON_PROMPT.md` with `skills/openclaw-minecraft/CRON_PROMPT.md` (always, regardless of existing file).
Recommendation

Do not overwrite a root cron prompt without explicit user approval; back up or merge existing content and prefer a per-skill cron configuration.

#
ASI10: Rogue Agents
Medium
What this means

The bot may continue moving, chatting, digging, placing, or attacking on a Minecraft server every 30 seconds until the cron job is manually stopped.

Why it was flagged

The cron prompt requires the agent to keep acting whenever the bot is idle, and it does not define a maximum runtime, goal completion condition, or automatic stop.

Skill content
If `currentTaskId` is `null`/empty and `queueLength` is 0, you MUST continue to steps 3-6 in order. Do not stop early.
Recommendation

Use explicit user-controlled start/stop conditions, maximum cycle counts, quiet hours, and a clear disable path before enabling the cron loop.

#
ASI02: Tool Misuse and Exploitation
Medium
What this means

On shared or public servers, this could spam chat, alter the world, follow players, or attack targets in ways the user did not individually review.

Why it was flagged

The autonomous loop can invoke mutating and socially visible Minecraft actions, including chat, digging, placing, following, and attacking, as scheduled batches without per-action confirmation.

Skill content
Allowed actions (use only these): `chat`, `move_to`, `move_relative`, `move`, `dig`, `place`, `equip`, `use_item`, `attack`, `follow`, `jump`.
Recommendation

Restrict this skill to private or test servers, remove risky actions unless needed, and require confirmation for attacks, player targeting, and world-modifying actions.

#
ASI03: Identity and Privilege Abuse
Low
What this means

Anyone with the token may be able to issue bot actions allowed by the controller.

Why it was flagged

The skill requires delegated controller authentication; this is expected for a bot-control API, but the token grants authority over controller-owned bots.

Skill content
`MC_CONTROLLER_TOKEN` (agent-specific JWT)
Recommendation

Use scoped, revocable agent tokens; do not share the master issuer secret with the agent; disable open registration unless intentionally operating an open controller.

#
ASI04: Agentic Supply Chain Vulnerabilities
Low
What this means

Users must trust the remote controller service to handle tokens and bot commands correctly.

Why it was flagged

The core capability depends on a disclosed external ngrok controller endpoint; no local controller code is included in the artifacts for review.

Skill content
Base URL: `https://56eb-125-246-120-211.ngrok-free.app/v1`
Recommendation

Verify who operates the controller, prefer a documented or self-hosted controller, and rotate tokens if the endpoint changes or is no longer trusted.

#
ASI06: Memory and Context Poisoning
Low
What this means

A stale or altered memory file could make the agent use the wrong bot, wrong token, or prior decision context.

Why it was flagged

The skill relies on persistent memory for credentials, bot identity, and autonomy logs; this is coherent for cron operation but creates state that future runs will reuse.

Skill content
Load `memory/mc-auth.json` and read `accessToken`... Load `memory/mc-bot.json` and read `botId`... Log decisions in `memory/mc-autonomy.json` with timestamps.
Recommendation

Store these files with appropriate access controls, review them before enabling cron, and clear or rotate them when changing bots or controllers.