Back to skill

Security audit

jovay-dapp

Security checks for vulnerabilities and agentic risk

Overview

The skill is coherent for building Jovay dApps, but it asks users to run mutable installs/templates and handle wallet secrets in exposed command forms.

Install only if you are comfortable with npm-based tool installation, remote template scaffolding, and local blockchain deployment commands. Use a fresh low-value testnet wallet, avoid putting real private keys or encryption keys directly into shell commands, inspect the generated template and package scripts before running installs, and pin versions where possible.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:10
Finding

Unpinned Third-Party Dependencies Create a Supply-Chain Execution Risk

Content
View full analysis
# Install base deps (Hardhat, ethers, etc.) npm install # Install OpenZeppelin contracts (almost always needed) npm install @openzeppelin/contracts # If the template has an install script: # chmod +x tools/install_deps.sh && ./tools/install_deps.sh ``` ### Technical Analysis The Skill installs `@jovaylabs/jovay-cli` and `@openzeppelin/contracts` without specifying exact reviewed versions. It also runs a general `npm install`, whose resolved dependency graph depends on the generated project's manifest and lockfile state. npm packages can execute lifecycle scripts during installation. If an upstream package, maintainer account, transitive dependency, or published version is compromised, installation can execute attacker-controlled JavaScript with the permissions of the user running the Agent. Installing the Jovay CLI globally increases the scope because its executable is placed in a shared command path and may subsequently be trusted by other projects or sessions. The Skill does not require a committed lockfile, `npm ci`, integrity verification, provenance checks, or suppression and review of lifecycle scripts before installation. ### Attack Path 1. An attacker compromises an upstream npm maintainer account, package release process, or transitive dependency. 2. The attacker publishes a malicious version or modifies a dependency selected by an unpinned version range. 3. The Agent follows the Skill and executes `npm install -g @jovaylabs/jovay-cli`, `npm ...[truncated 974 chars]
Remediation
View remediation

T03 · Remote Payload Retrieval and Execution

Warning
Location
SKILL.md:127
Finding

Mutable Remote Template Can Introduce and Execute Unreviewed Code

Content
View full analysis
``` ```markdown This clones the `EasyJovayDappTemplate` from GitHub and creates the project structure: ``` ```bash cd # Install base deps (Hardhat, ethers, etc.) npm install # Install OpenZeppelin contracts (almost always needed) npm install @openzeppelin/contracts # If the template has an install script: # chmod +x tools/install_deps.sh && ./tools/install_deps.sh ``` ### Technical Analysis The initialization command clones `EasyJovayDappTemplate` from GitHub, but the Skill does not require the template to be pinned to a reviewed commit, tag with verified signature, or cryptographic digest. Consequently, the files retrieved during a future invocation may differ from the files that existed when the Skill was audited. The instructions subsequently allow the Agent to mark `tools/install_deps.sh` executable and run it. That shell script originates from the retrieved template and can execute arbitrary operating-system commands. Running `npm install` in the retrieved project can additionally execute package lifecycle scripts defined by the remote template. The shell command is presented as optional and commented, so execution is not guaranteed. Nevertheless, the documented workflow explicitly establishes an execution path from mutable remote content to local command execution without an intervening authenticity or code-review requirement. ### Attack Path 1. An attacker compromises the template repository, its maintainer credentials, the template-selection mechanism in the Jovay CLI, or another component controlling the retrieved revision. 2. The attacker modifies `tools/install_deps.sh`, `package.json`, or another initialization artifact. 3. The Agent executes `jovay dapp init --name
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:791
Finding

Wallet Secrets Are Passed Through Exposed Command Invocation Channels

Content
View full analysis
``` ```bash jovay dapp deploy --enc-key ``` ```bash JOVAY_TESTNET_RPC_URL="" DEPLOYER_PRIVATE_KEY="" \ npx hardhat run scripts/deploy.js --network jovay_testnet ``` ```markdown 9. **Encrypted wallet**: If the user set up an encrypted wallet with `jovay init --enc`, they need `--enc-key` for build and deploy commands. ``` ### Technical Analysis The wallet encryption key is supplied through the `--enc-key` command-line argument. Command-line arguments may be observable through process-inspection interfaces while the command is running and can be captured by shell history, terminal logging, process-monitoring agents, audit systems, or diagnostic tooling. The Hardhat example encourages users to substitute the deployment private key directly into an interactive shell command as an inline environment assignment. Although the private key is not an argument to Hardhat, entering its literal value can still persist it in shell history, copied command logs, terminal recordings, CI output, or support transcripts. On some systems, process environment data may also be accessible to sufficiently privileged or same-context monitoring processes. These values protect wallet signing authority. Treating them as ordinary command text creates unnecessary exposure compared with protected secret files, secret managers, or hidden interactive input. ### Attack Path 1. A user replaces `` or `` with a real secret and executes the documented command. 2. The encryption key becomes part of the command argument vector, or the private key becomes literal shell command text. 3. A local process monitor, t ...[truncated 1275 chars]
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (9)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 147)May include surrounding context.

├── hardhat.config.cjs # Hardhat configuration ├── package.json # Root package.json with Hardhat deps └── .jovay/ └── .env # JSON: { PROJECT_NAME, TEMPLATE_NAME }

text

### 2.3 Install Dependencies

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The instruction says to act when 'the user says they want to generate a dApp,' but it does not define specific trigger phrases, scope boundaries, or exclusion conditions. Because this is a markdown skill file, that broad natural-language activation condition could overlap with many ordinary requests about discussing or learning about dApps rather than actually invoking full project generation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The example query 'Help me create a token airdrop dApp' uses a very broad lead-in ('Help me create') that is common conversational language. In a markdown skill description, presenting such generic wording as an invocation example without tighter constraints can encourage unintended activation on routine assistance requests.

Content

No source excerpt is available for this finding.