Back to skill

Security audit

CVE Audit Skill

Security checks for vulnerabilities and agentic risk

Overview

The skill has a legitimate dependency-auditing purpose, but it repeatedly tells agents to run an unpinned npm package with broad local access.

Review before installing. If you use this skill, pin osv-ui to a specific reviewed version or install it through a lockfile, run it with limited filesystem and environment access, and only invoke it for explicit dependency-audit requests. Confirm generated package changes before applying them.

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

Error
Location
SKILL.md:30
Finding

Unpinned Third-Party Package Execution Through npx

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 30-36; repeated at lines 49, 65, and 77-81
Vulnerability Type: Unpinned third-party dependency execution
Risk Level: High

Vulnerable Code

bash
# Single service
npx osv-ui --no-open --json ./osv-report.json

# Multiple services
npx osv-ui ./frontend ./api ./worker --no-open --json ./osv-report.json

# Auto-discover
npx osv-ui --discover --no-open --json ./osv-report.json

Additional unpinned invocations include:

bash
npx osv-ui [same paths]
npx osv-ui --no-open --json ./osv-report-after.json
npx osv-ui
npx osv-ui ./frontend ./api
npx osv-ui --discover
npx osv-ui --json=report.json --no-open
npx osv-ui --html=report.html --no-open

Technical Analysis

The Skill instructs the Agent to execute osv-ui through npx without specifying an exact package version or verifying package integrity. If the package is not already available locally, npx can retrieve and execute the current package release from the configured npm registry.

Consequently, the code executed during an audit is mutable and may differ from the code reviewed when this Skill was published. The project contains no vendored implementation, package lockfile, integrity hash, or other mechanism that binds execution to a specific audited artifact. The confirmation requirement in the Skill applies only to dependency fixes and does not protect the initial npx execution.

This creates a supply-chain execution boundary: compromise of the upstream package, npm publisher account, registry resolution, or local registry configuration could cause attacker-controlled JavaScript to execute with the permissions of the Agent or user.

Attack Path

  1. An attacker compromises the upstream osv-ui package, its publisher account, or a registry/mirror used by the environment.
  2. The attacker publishes or serves a malicious package release under the expected package ...[truncated 1349 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin osv-ui to a reviewed, exact version in every command rather than resolving the latest release:

    bash
    npx --yes osv-ui@1.0.0 --no-open --json ./osv-report.json
    
  2. Install the pinned dependency through a committed lockfile and use a deterministic installation method such as npm ci. Review and preserve the registry-provided integrity value in the lockfile.

  3. Verify the package's publisher, provenance, release signatures where available, repository correspondence, and npm integrity metadata before trusting it.

  4. Require explicit user confirmation before downloading or executing the package, rather than requiring confirmation only before applying dependency upgrades.

  5. Run the scanner in a restricted environment with:

    • Read-only access to the project where practical.
    • No access to unrelated home-directory files or credential stores.
    • Minimal environment variables.
    • Restricted outbound network access.
    • No elevated operating-system privileges.
  6. Avoid globally configured or untrusted npm registries and mirrors. Explicitly use a trusted registry and enforce dependency allowlisting where supported.

  7. Update every repeated invocation in SKILL.md, including the dashboard, verification, and quick-reference commands, so none can silently fall back to an unpinned package release.

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • 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
Findings (11)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The activation triggers are broad enough to match general discussion of security, CVEs, or package health even when the user has not clearly requested this specific skill. In an agent setting, that can lead to unnecessary execution of external tooling and browser-opening behavior, which increases exposure to the unpinned-package risks elsewhere in the skill.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The skill repeatedly instructs use of npx osv-ui without a pinned package version, which causes code to be fetched and executed from the npm registry at run time. If the package is updated maliciously, hijacked, or affected by a compromised dependency chain, the agent could execute attacker-controlled code during a security audit workflow.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

This command invokes npx osv-ui without version pinning, so execution depends on whatever package version npm resolves at the time. That creates a supply-chain execution path where a malicious or unexpectedly changed release can run arbitrary code in the user's environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx osv-ui --discover without pinning both downloads and executes remote package code while also scanning auto-discovered directories. The unpinned execution is the core issue: a compromised upstream package could gain code execution in a context likely to have source access and developer credentials.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

The dashboard step again relies on npx osv-ui with no version constraint, creating the same runtime supply-chain risk. Because this step also opens a local web dashboard, it may encourage interactive use of a package whose exact code version is unknown at execution time.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The verification re-scan still uses an unpinned npx package, so even the safety-checking phase can execute untrusted newly resolved code. Security tooling is especially sensitive because users are more likely to grant it broad filesystem and network access.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The quick-reference command npx osv-ui is an unpinned package execution pattern and therefore exposes users to arbitrary code changes from upstream. Quick-reference snippets are likely to be copied verbatim, increasing the chance the unsafe pattern is propagated into real use.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

This multi-service quick-reference example repeats the same unpinned npx execution risk. In practice it may be run from monorepos or privileged development environments, making compromise more consequential.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The auto-detect quick-reference command executes whatever npx resolves for osv-ui, which is unsafe for a security-focused skill. Combining unpinned execution with repository discovery broadens the practical blast radius if the package is malicious.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The JSON export example still fetches and runs an unpinned npm package at execution time. Because this is instructional documentation, the unsafe pattern may be reused widely and undermines trust in the security-auditing workflow.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

This HTML export example also executes an unpinned package from npm, with the same supply-chain risk. Since the skill is framed as a security tool, users may trust and run it in sensitive environments, increasing potential impact.

Content

No source excerpt is available for this finding.