Back to skill

Security audit

pruna

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent Pruna media-suite installer skill with disclosed API/tool requirements and no evidence of hidden, destructive, persistent, or deceptive behavior.

Install only if you trust the PrunaAI skill source and are comfortable configuring Pruna and, for audio tools, Replicate credentials. Treat media generation as potentially billable, review API usage before bulk workflows, and prefer pinned or verified installer versions where possible.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (33)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill bundles many tools and workflows that rely on external Pruna and Replicate APIs, but the description does not clearly warn users that use may trigger paid API calls and charges. This is dangerous because users may unknowingly incur financial cost, especially when the skill encourages multi-step generation workflows and bulk media operations.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The installation command uses npx skills without pinning the CLI/package version, so execution depends on whatever version is current at install time. That creates a supply-chain risk: a compromised or breaking upstream release could run unexpected code during install or change behavior without review.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

This command references npx skills without a pinned version, meaning users may execute a different CLI release over time. If the upstream package is maliciously updated or introduces insecure behavior, the skill documentation becomes an implicit code-execution vector.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

Using npx skills without version pinning allows mutable upstream code to be fetched and executed at runtime. For installation tooling, that is a real supply-chain exposure because documentation instructs users to trust the latest published package automatically.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The documented npx skills usage is unpinned, so future installs may run different code than the author originally validated. This increases exposure to dependency hijacking, account compromise of the publisher, or accidental insecure releases.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

Because npx resolves and runs the latest matching package by default, the command is vulnerable to upstream package drift. In security-sensitive contexts, install docs that cause arbitrary newer code execution are a meaningful supply-chain weakness.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

This unpinned npx skills reference can cause users to execute code that has not been reviewed in the context of this skill. If the package changes or is compromised, the documented workflow can expose systems or secrets during installation.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The skill documentation directs users to an unversioned npx command, which is a common supply-chain risk pattern. It permits silent changes in installed behavior and expands the trust boundary to all future upstream releases.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

An unpinned npx skills invocation means this documentation will execute mutable remote code whenever followed. That is dangerous because the referenced package can change independently of the skill and could be abused for malicious installation-time execution.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The command depends on npx resolving the current package version, which undermines reproducibility and increases supply-chain risk. Users may unknowingly run a compromised or incompatible release simply by following the docs.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

Referencing npx skills without a fixed version allows the executable code path to change over time without any update to this skill file. That creates a real risk of malicious or accidental unsafe behavior being introduced upstream.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The documentation instructs execution of an unpinned remote package via npx, which is a classic supply-chain weakness. If the package publisher or dependency tree is compromised, users can be exposed immediately during install.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

Because the skills package version is not fixed, the exact code executed by npx is nondeterministic over time. That lack of pinning makes the install path vulnerable to malicious updates, typosquatting-like confusion, or breaking changes with security consequences.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

This installation example executes a package resolved at runtime rather than a reviewed, fixed artifact. In practice, that increases the risk of arbitrary code execution through compromised upstream publishing or unauthorized package changes.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The unpinned npx skills reference expands trust to future unpublished-at-review-time versions of the CLI. That is dangerous because the command may later execute code with different security properties than intended by the skill author.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

By using npx without a version constraint, this command makes installs non-reproducible and exposes users to supply-chain compromise. A later upstream release could inject malicious behavior while the skill file appears unchanged.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The skill's install guidance triggers execution of a mutable remote package through npx, which is unsafe from a supply-chain perspective. If abused, an attacker could run arbitrary code on systems that follow the documentation.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The npx skills command is not pinned to a reviewed version, so the actual executable can change independently of this skill release. That makes the documentation itself a conduit for unreviewed code execution and supply-chain attacks.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

Unpinned npx execution allows future package versions to alter installation behavior without any visible change in the docs. In a skill that installs many components, this broadens the blast radius of any upstream compromise.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

Because the command relies on the latest skills package, it is susceptible to malicious or accidental upstream changes. This is a genuine vulnerability pattern in documentation that asks users to execute remote tooling locally.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

This npx skills invocation is unversioned, so the command's behavior is not stable or reviewable over time. Attackers who compromise the package distribution path could exploit that to achieve code execution on user machines.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

Unpinned package execution through npx is dangerous because it trusts whatever current artifact the registry serves. That creates a supply-chain attack window and weakens reproducibility for users following the installation steps.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The command example does not constrain the version of the tool being executed, so users may run code that the skill author never reviewed. This is especially concerning for setup/install commands, which often execute with local user privileges and access to tokens.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

Using npx skills without pinning introduces a direct supply-chain risk because the remote executable can change at any time. The danger is elevated by the surrounding context, which discusses API keys and media workflows likely to run on developer workstations with access to secrets and files.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

This install instruction fetches and runs a mutable package via npx, leaving users exposed to malicious upstream changes or compromised publisher accounts. Since the skill aggregates many sub-skills, a single compromised install path could affect the entire suite deployment.

Content

No source excerpt is available for this finding.