Back to skill

Security audit

Solanaprox Ai

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent Solana-paid AI gateway, but it uses a public wallet address as the payment credential and exposes broad paid orchestration without clear spending or action controls.

Review this carefully before installing. Use a wallet with only limited funds, confirm whether solanaprox.com requires proof of wallet ownership beyond the address, avoid sending secrets or regulated data in prompts, and pin or audit the npm package if using the SDK.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:69
Finding
Unpinned Third-Party npm Dependency## Vulnerability Details **File Location**: `SKILL.md`, lines 69–72 **Vulnerability Type**: Unpinned third-party package installation **Risk Level**: Medium ### Vulnerable Code ```bash npm install solanaprox-openai ``` ### Technical Analysis The documented installation command does not specify an exact package version or verify package integrity. It therefore resolves the mutable release associated with npm's default distribution tag at installation time. The project contains no copy of the package source, lockfile, checksum, signature, or provenance verification instructions. Consequently, the code eventually installed and executed may differ from what was reviewed. npm packages can also define lifecycle scripts that execute during installation. This creates a supply-chain risk if the package, its maintainer account, or any transitive dependency is compromised. The finding does not establish that `solanaprox-openai` is currently malicious; it identifies the unsafe dependency-installation practice. ### Attack Path 1. An attacker compromises the package publisher, the package itself, or a transitive dependency. 2. The attacker publishes a malicious release under the package's default distribution tag. 3. A user follows the documented `npm install solanaprox-openai` instruction. 4. npm resolves and downloads the attacker-controlled release. 5. Malicious lifecycle scripts may run during installation, or malicious package code may run when the documented SDK is imported and used. 6. That code executes with the permissions of the user or process performing the installation or running the application. ### Impact Assessment Successful exploitation could permit arbitrary code execution within the installing user's privilege boundary. Depending on the environment, this may expose application data, environment variables, wallet addresses, prompts, accessible credentials, and files writable by that account. It may also al ...[truncated 299 chars]
Remediation
## Remediation Suggestions 1. Pin the dependency to an exact, reviewed version rather than relying on npm's mutable default tag: ```bash npm install --save-exact solanaprox-openai@<reviewed-version> ``` 2. Commit and enforce an npm lockfile in consuming projects, and use `npm ci` for reproducible installation. 3. Verify npm package provenance, publisher identity, and integrity metadata before installation. 4. Audit the package, its lifecycle scripts, and its complete transitive dependency tree. 5. During initial verification, disable lifecycle scripts where compatible: ```bash npm install --ignore-scripts --save-exact solanaprox-openai@<reviewed-version> ``` 6. Run dependency installation and application execution with least privilege in an isolated environment without unrelated credentials. 7. Establish an upgrade-review process so version changes require renewed source and dependency inspection.
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 (3)

External Transmission

Medium
Category
Data Exfiltration
Content
# 2. Deposit USDC to your wallet's SolanaProx balance
# 3. Use wallet address as credential

curl -X POST https://solanaprox.com/v1/messages \
  -H "Content-Type: application/json" \
  -H "X-Wallet-Address: $SOLANAPROX_WALLET_ADDRESS" \
  -d '{
Confidence
88% confidence
Finding
This skill explicitly transmits user-supplied content and the wallet address in an HTTP request to an external service. In context, external transmission is the core function of the skill, but it still creates real privacy and financial-risk exposure because prompts may contain sensitive data and the wallet address is used as a billing credential.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill shows example requests that send prompts and the user's wallet address to solanaprox.com, but it does not clearly warn users that both their input data and wallet identifier are transmitted to a third-party service for inference and billing. This can lead users to disclose sensitive prompts or use a wallet-linked identity without understanding the privacy and tracking implications.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill describes orchestration, spending metadata, and wallet-based payment, but it does not clearly warn that invoking the service can consume pre-deposited USDC and may trigger additional paid actions through multi-agent workflows. Users may unintentionally incur financial cost, especially when orchestration expands a single request into multiple backend operations.

Static analysis

No suspicious patterns detected.