Skill Operator

Security checks across malware telemetry and agentic risk

Overview

This package is advertised as a skill manager, but the supplied artifacts implement agent-to-agent messaging and file transfer with network services and auto-installed dependencies.

Do not install this as 'Skill Operator.' Only consider it if you intentionally want the AgentComm messaging/file-transfer skill, have verified the source, pinned dependencies, restricted file access, protected the Nostr private key, and limited LAN use to trusted networks.

VirusTotal

1/64 vendors flagged this skill as malicious, and 63/64 flagged it 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.

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

A user installing this as a skill manager would instead give the agent network messaging and file-transfer capabilities.

Why it was flagged

The actual skill instructions describe a decentralized communication and file-sharing system, not the registry's advertised 'Skill Operator' skill-management purpose.

Skill content
# AgentComm

**Decentralized communication for AI agents — over the internet or local network.**
Recommendation

Do not install it under the Skill Operator listing; treat it as a mislabeled package unless the publisher corrects the metadata and provenance.

#
ASI02: Tool Misuse and Exploitation
High
What this means

The agent could be directed to transmit local files outside the machine through the communication workflow.

Why it was flagged

The skill exposes arbitrary local file sharing to another public key, but the advertised package purpose is skill management and the artifacts do not define path limits or approval requirements.

Skill content
share_file file_path="/path/to/file.pdf" target_pubkey="npub1..."
Recommendation

Only use file-sharing tools with explicit user confirmation, narrow path allowlists, and clear disclosure that files may leave the local environment.

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

Other devices on a reachable network could send messages to the agent or read stored LAN messages.

Why it was flagged

The LAN server listens on all interfaces and exposes received messages over an unauthenticated /messages endpoint.

Skill content
HTTPServer(('0.0.0.0', self.port), self._create_handler()) ... elif self.path == "/messages": ... self.wfile.write(json.dumps(messages).encode())
Recommendation

Bind only to trusted interfaces, require authentication, and avoid using this on public or untrusted networks.

#
ASI05: Unexpected Code Execution
Medium
What this means

Installing or importing the skill can fetch and execute additional package code not captured by the declared installation metadata.

Why it was flagged

The code runs pip installation automatically when a dependency is missing, despite the registry saying there is no install spec.

Skill content
except ImportError:
    # Install nostr library
    subprocess.run(['pip', 'install', 'nostr'], check=True)
Recommendation

Move dependency installation to an explicit install spec, pin versions, and avoid import-time package installation.

#
ASI04: Agentic Supply Chain Vulnerabilities
Medium
What this means

Future or unexpected dependency versions could change the code that runs in the user's environment.

Why it was flagged

Dependencies are declared without version pins, and the code also installs them dynamically, creating a provenance and reproducibility gap.

Skill content
nostr
zeroconf
requests
Recommendation

Pin dependency versions and hashes, declare installation requirements in metadata, and verify package provenance before use.

#
ASI03: Identity and Privilege Abuse
Medium
What this means

A Nostr private key can decrypt and sign messages for the user's agent identity; mishandling it could expose or compromise that identity.

Why it was flagged

The skill handles a private identity key, but registry metadata declares no required environment variables or primary credential.

Skill content
- `NOSTR_PRIVATE_KEY`: Your Nostr private key (optional)
Recommendation

Declare this credential explicitly, store it securely, avoid returning it in normal tool output, and limit access to trusted workflows.