ClawHub
Back to skill

Security audit

Solana On-Chain Agent Registry, Reputation & Trust

Security checks across malware telemetry and agentic risk

Overview

This is a coherent Solana SDK guide, but users should treat wallet signing, IPFS uploads, and on-chain changes as real high-impact actions.

Install this only if you intend to use the 8004 Solana registry SDK. Start on devnet, use a dedicated low-balance wallet, pin and verify npm dependencies, review all IPFS/on-chain metadata for secrets or private details, and manually approve every transaction, delete, transfer, or immutable metadata action.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill instructs users to load a Solana private key directly from an environment variable and immediately construct a signer, but it does not warn that this key grants full control over on-chain assets and should never be logged, hardcoded, shared with untrusted tools, or used in broad-scope agent environments. In an agent-skill context, this is more dangerous than ordinary SDK docs because agents may be given ambient environment access and could inadvertently expose or misuse the key during automated workflows.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The documentation shows uploading metadata and related content to IPFS without warning that the data leaves the local system and may become publicly retrievable, persistent, and difficult to remove. In this skill, users are encouraged to upload agent metadata, feedback, and proofs, so omission of a privacy warning can lead to accidental disclosure of internal endpoints, business data, or user-linked records.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
These examples perform live blockchain write operations such as registering an agent and setting an operational wallet, which can incur fees and create durable on-chain state, yet they are presented without an explicit caution about cost, permanence, and authority implications. In an agent-executable skill, this increases risk because an automated system may treat the snippets as routine actions and trigger financially impactful or hard-to-reverse changes.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill includes examples for deleting metadata, transferring ownership, and syncing ownership state without a prominent caution, even though these actions can remove data, relinquish control, or alter authority relationships in ways that may be irreversible or operationally disruptive. This is especially risky in an agent skill because the commands are concise and actionable, making accidental execution by an automated agent more plausible than in passive reference documentation.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.