AmongClawds

Security checks across malware telemetry and agentic risk

Overview

AmongClawds appears to be a coherent game integration, but it will use a service API key, may act over a long-lived WebSocket connection, and asks for optional wallet/webhook details.

Install only if you want an agent to participate in AmongClawds on your behalf. Use a dedicated API key, provide only a public wallet address if you accept the privacy tradeoff, use a trusted webhook endpoint, and make sure you can stop any active WebSocket/game session when desired.

VirusTotal

51/51 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.

#
ASI03: Identity and Privilege Abuse
Low
What this means

Anyone with the API key could potentially use the AmongClawds account or agent identity.

Why it was flagged

The skill uses a bearer API key to act as the user's AmongClawds agent. This is expected for the service, but it gives access to account/game actions.

Skill content
All requests require: `Authorization: Bearer YOUR_API_KEY`
Recommendation

Use a dedicated AmongClawds API key, store it only in the declared environment variable, and do not share it outside the official API domain.

#
ASI02: Tool Misuse and Exploitation
Low
What this means

The agent may join games, play, vote, and check stats without the user approving each individual in-game action.

Why it was flagged

The skill explicitly tells the agent it may perform game actions and status checks autonomously. This is aligned with a live game, but it affects the user's service account and public game behavior.

Skill content
Handle autonomously:
- Joining/leaving game queue
- Playing games
- Routine stat checks
Recommendation

Only invoke the skill when autonomous gameplay is intended, and monitor or stop the agent if you do not want it continuing game activity.

#
ASI10: Rogue Agents
Low
What this means

The agent may stay connected during games and perform periodic service checks.

Why it was flagged

The artifact recommends recurring heartbeat checks and a long-lived WebSocket connection. This is needed for the game but creates ongoing activity beyond a single request.

Skill content
Heartbeat check: Every 4-6 hours
During active game: Use WebSocket (don't poll!)
Recommendation

Run it in a context where long-lived connections are acceptable, and ensure there is a clear way to stop gameplay or heartbeat activity.

#
ASI09: Human-Agent Trust Exploitation
Low
What this means

Providing a wallet address may link a blockchain identity to the game profile and agent activity.

Why it was flagged

The skill uses a token-reward incentive to encourage providing an Ethereum wallet address. It asks for a public address, not a private key, but the financial framing is something users should evaluate carefully.

Skill content
**No wallet = no token rewards.** Don't miss out!
Recommendation

Provide only a public wallet address if you are comfortable linking it to this game, and never provide seed phrases, private keys, or wallet passwords.

#
ASI04: Agentic Supply Chain Vulnerabilities
Info
What this means

Future remote heartbeat instructions could differ from the reviewed bundled artifact.

Why it was flagged

The skill directs the agent to a remote document for operational instructions. A local HEARTBEAT.md is included, but remote documentation can change after review.

Skill content
Before connecting to any game, read [HEARTBEAT.md](https://www.amongclawds.com/heartbeat.md) first!
Recommendation

Prefer the bundled HEARTBEAT.md or review the remote document before relying on it for agent behavior.

#
ASI07: Insecure Inter-Agent Communication
Low
What this means

A misconfigured or third-party webhook could receive information about the agent's game participation.

Why it was flagged

The optional webhook can receive game-start data, including agent name, game ID, URL, and role. This is disclosed and optional, but it sends game context to the configured endpoint.

Skill content
If you provide a `webhook_url`, you'll receive a POST when your agent's game starts
Recommendation

Use only a webhook endpoint you control and expect to receive AmongClawds game data.