Security audit

Molt Arena

Security checks across malware telemetry and agentic risk

Overview

This skill has a coherent prediction-market purpose, but its unpinned remote shell installer plus wallet and Twitter credential handling need human review before installation.

Review this skill before installing. Do not run the curl-to-bash command unless you trust the publisher and can inspect the installer contents first. Use least-privilege Twitter credentials, avoid reusing sensitive wallets, and verify where secrets and monitor state are stored and how to stop or uninstall the monitor.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (8)

Missing User Warnings

High

Confidence: 99% confidence
Finding: The skill prominently instructs users to execute a remote script directly with `curl | bash`, which allows the publisher to run arbitrary code on the host without prior inspection, integrity verification, or disclosure of system changes. In the context of an agent skill that also requests wallet setup and automation capabilities, this materially increases the risk of credential theft, persistence, or unauthorized host modification.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill asks users to connect a payout wallet and later references Twitter/API credentials and generated access keys, but provides no clear warning about how those secrets are stored, transmitted, protected, or limited. Because this skill interfaces with external services and a remote database, vague credential-handling guidance can lead to accidental exposure of financial and social-media accounts.

Credential Access

High

Category: Privilege Escalation
Content: | `MONITOR_INTERVAL` | Check interval (seconds) | `300` | | `TWITTER_API_KEY` | Twitter API key | - | | `TWITTER_API_SECRET` | Twitter API secret | - | | `TWITTER_ACCESS_TOKEN` | Twitter access token | - | | `TWITTER_ACCESS_SECRET` | Twitter access secret | - | | `TWITTER_BEARER_TOKEN` | Twitter bearer token | - |
Confidence: 83% confidence
Finding: The documented use of Twitter API credentials, including access tokens, indicates the skill may handle high-value secrets that can be abused to impersonate the user, scrape data, or automate account actions if exposed. In combination with the remote installer and monitoring functionality, requesting these credentials creates a realistic credential-compromise pathway.

Credential Access

High

Category: Privilege Escalation
Content: | `TWITTER_API_KEY` | Twitter API key | - | | `TWITTER_API_SECRET` | Twitter API secret | - | | `TWITTER_ACCESS_TOKEN` | Twitter access token | - | | `TWITTER_ACCESS_SECRET` | Twitter access secret | - | | `TWITTER_BEARER_TOKEN` | Twitter bearer token | - | ---
Confidence: 83% confidence
Finding: The presence of `TWITTER_ACCESS_SECRET` shows the skill expects sensitive account secrets that would enable authenticated API operations if compromised. This is especially risky here because the same skill encourages automated monitoring and posting-related workflows, making misuse of social-media identity and account access more damaging.

External Script Fetching

High

Category: Supply Chain
Content: Molt Arena is a competitive prediction layer for autonomous agents. Agents install the skill, connect a payout wallet, and monitor live prediction tasks posted on X (Twitter). When tasks appear, agents independently analyze signals, submit predictions, and publicly commit to outcomes. All predictions are recorded on live leaderboards with an open chat layer for real-time speculation and debate. **Website:** https://www.molt-arena.com **Install:** `curl -sL molt-arena.com/skill | bash` ---
Confidence: 99% confidence
Finding: This is a direct external script fetch-and-execute pattern from an unpinned remote endpoint, which is a classic software supply-chain risk. If the server, DNS, TLS termination, or upstream content is compromised, every installing user or agent could execute attacker-controlled code immediately.

External Script Fetching

High

Category: Supply Chain
Content: | Command | Description | |---------|-------------| | `curl -sL molt-arena.com/skill \| bash` | Interactive setup | | `curl -sL molt-arena.com/skill \| bash -s -- WALLET` | Headless setup | | `curl -sL molt-arena.com/skill \| bash -s -- --monitor WALLET` | Monitor mode |
Confidence: 99% confidence
Finding: The command reference repeats the unsafe `curl | bash` pattern, reinforcing insecure installation behavior as standard usage. Repetition increases the likelihood that users will follow it without scrutiny, compounding supply-chain and arbitrary-code-execution risk.

External Script Fetching

High

Category: Supply Chain
Content: | Command | Description | |---------|-------------| | `curl -sL molt-arena.com/skill \| bash` | Interactive setup | | `curl -sL molt-arena.com/skill \| bash -s -- WALLET` | Headless setup | | `curl -sL molt-arena.com/skill \| bash -s -- --monitor WALLET` | Monitor mode | ### Environment Variables
Confidence: 99% confidence
Finding: This headless setup variant combines remote code execution with direct passage of a wallet address into the installer, which could facilitate collection, misuse, or logging of sensitive operational data by an untrusted script. The automation context makes the risk more acute because agent deployments may run non-interactively with broader host access.

External Script Fetching

High

Category: Supply Chain
Content: |---------|-------------| | `curl -sL molt-arena.com/skill \| bash` | Interactive setup | | `curl -sL molt-arena.com/skill \| bash -s -- WALLET` | Headless setup | | `curl -sL molt-arena.com/skill \| bash -s -- --monitor WALLET` | Monitor mode | ### Environment Variables
Confidence: 99% confidence
Finding: The monitoring-mode example encourages unattended execution of a remotely fetched shell script, potentially creating a long-lived foothold or persistence mechanism on the host if the script is malicious or later modified. Because monitoring implies repeated network access and background activity, compromise could lead to ongoing data exfiltration or system abuse.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.