Back to skill

Security audit

Unraid CLI

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent Unraid administration CLI, but it documents handling powerful API keys over plaintext HTTP and installs an unpinned external npm CLI that would receive those credentials.

Review this carefully before installing. Use HTTPS with certificate validation, avoid passing API keys on the command line, store any config file with owner-only permissions, use a dedicated least-privilege Unraid API key, and prefer a pinned or otherwise verified CLI release before giving it access to your server.

Vulnerability Patterns
  • 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
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:104
Finding

API Key and Administrative Data Transmitted over Plaintext HTTP

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:104-109; related examples also appear in README.md:39 and README.md:47-52
Vulnerability Type: Plaintext transmission of sensitive credentials and server data
Risk Level: High

Vulnerable code snippet (SKILL.md:104-109):

yaml
default_profile: lab
profiles:
  lab:
    host: http://192.168.1.10:7777
    apiKey: YOUR_API_KEY
    output: json

Related command example (README.md:39):

bash
ucli --host http://192.168.1.10:7777 --api-key YOUR_API_KEY system info

Related configuration example (README.md:47-52):

yaml
default_profile: lab
profiles:
  lab:
    host: http://192.168.1.10:7777
    apiKey: YOUR_API_KEY
    output: table
    timeout: 30

Technical Analysis

The documented configuration sends an API key to an Unraid GraphQL endpoint using unencrypted HTTP. The network transmission is necessary for the Skill's declared remote-management functionality, but plaintext transport is not the minimum safe mechanism for carrying administrative credentials.

An attacker able to observe or manipulate traffic on the network path may recover the API key, inspect sensitive GraphQL responses, or modify requests and responses. The affected data can include server health information, disk and share details, network configuration, container and VM metadata, logs, notifications, and diagnostic results.

The command-line example also places the API key in a process argument. Depending on the operating system and shell configuration, this may expose the credential through process listings, command history, diagnostic collection, or terminal logs.

Attack Path

  1. A user or agent follows the documented example and connects to the Unraid API through http://.
  2. An attacker gains visibility into the network path, such as through a compromised local device, malicious access point, switched-network at ...[truncated 1176 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace all plaintext HTTP examples with HTTPS URLs and require TLS for authenticated API requests.
  2. Reject http:// endpoints by default. If plaintext transport is needed for isolated development, require an explicit insecure override accompanied by a prominent warning.
  3. Validate server certificates and hostnames. Do not recommend disabling certificate verification.
  4. For private or self-signed deployments, document how to configure a trusted local certificate authority rather than bypassing validation.
  5. Recommend dedicated, least-privilege API keys limited to the operations required by each automation task.
  6. Avoid passing API keys through --api-key where possible. Prefer a protected environment variable, restricted credential file, or operating-system secret store.
  7. Ensure configuration files containing credentials are created with owner-only permissions and are excluded from source control, logs, and diagnostics.
  8. Redact authentication material from verbose output, errors, telemetry, and diagnostic commands.
  9. Provide key rotation and revocation guidance in case a credential has previously been sent over plaintext HTTP.

T08 · Insecure Dependencies

Warning
Location
SKILL.md:14
Finding

Unpinned Third-Party CLI Installation Executes Unreviewed Package Releases

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:14-20; related installation declarations appear in README.md:29 and marketplace.json:15-19
Vulnerability Type: Unpinned external dependency and supply-chain exposure
Risk Level: Medium

Vulnerable code snippet (SKILL.md:14-20):

yaml
install:
  - id: npm
    kind: node
    package: unraid-cli
    bins:
      - ucli
    label: "Install unraid-cli (npm)"

Related installation command (README.md:29):

bash
npm install -g unraid-cli

Related marketplace declaration (marketplace.json:15-19):

json
"install": {
  "type": "npm",
  "package": "unraid-cli",
  "binary": "ucli"
},

Technical Analysis

The Skill installs unraid-cli without an exact version or integrity value. Package resolution can therefore select a newer npm release than the one intended or reviewed when the Skill was published. The artifact contains only documentation and metadata; it does not include the CLI source, a lockfile, a package archive hash, or other material sufficient to audit the installed executable.

This is particularly sensitive because the installed CLI receives the Unraid API key, communicates with the managed server, reads operational data, and can invoke administrative mutations. npm packages may also execute lifecycle scripts during installation. No evidence establishes that the current package is malicious; the vulnerability is the absence of controls that bind installation to a specific audited artifact.

The version metadata is also inconsistent: SKILL.md declares Skill version 0.4.0, while marketplace.json declares version 1.0.0. Although these may represent marketplace metadata rather than the npm package version, the discrepancy makes it harder to determine which release was intended and reviewed.

Attack Path

  1. An attacker compromises the npm publisher account, package publication pip ...[truncated 1282 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin the dependency to an exact audited npm version rather than resolving the latest release.
  2. Record and verify the package integrity digest or signed provenance before installation.
  3. Keep the Skill metadata, marketplace version, documented installation command, and audited package release synchronized.
  4. Include the executable source or a reproducible source reference so reviewers can verify the code that handles credentials and administrative requests.
  5. Use npm provenance and package-signing controls where available, and document how consumers can verify them.
  6. Prefer a project-local or isolated installation over a global installation where operationally practical.
  7. Review package lifecycle scripts and disable them during installation when they are not required.
  8. Run the CLI under a minimally privileged local account and provide it with a dedicated least-privilege Unraid API key.
  9. Establish an update process in which each new dependency version is reviewed, tested, pinned, and released deliberately rather than consumed automatically.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The README instructs users to pass API keys on the command line and store them in a local config file, but it does not warn about common exposure paths such as shell history, process listings, shared terminals, backups, or overly permissive file permissions. In an admin CLI for Unraid, those credentials likely grant broad control over infrastructure, so missing guidance meaningfully increases the chance of credential leakage.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · README.md (reported line 52)May include surrounding context.

Config file setup

Create ~/.config/ucli/config.yaml:

yaml
default_profile: lab

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
85% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · README.md (reported line 138)May include surrounding context.

md
| `--page`, `--page-size`, `--all` | Pagination |
| `--timeout`, `--retry` | Network settings |
| `--quiet`, `--verbose`, `--debug` | Verbosity |
| `--yes`, `--force` | Skip confirmations |

Environment variables supported for host, API key, profile, and config path. Check `ucli diagnostics env` for the exact resolution.