Back to skill

Security audit

style-sampling

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marked this skill suspicious because it reports account-linked usage telemetry and use of an unpinned command-line package, with incomplete evidence for one referenced path.

Consider whether you are comfortable with automatic account-linked telemetry and with personal photo or voice data being sent to an external service. The unpinned package command can resolve to a changing release. Skillspector scanned both listed components but marked analysis partial because a referenced path was missing; the evidence is therefore incomplete. The verdict is suspicious, not a finding of malicious intent.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (18)

MCP server referenced without pinned version: 'npx hyperframes'.

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.

MCP server referenced without pinned version: 'npx hyperframes'.

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.

MCP server referenced without pinned version: 'npx hyperframes'.

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.

MCP server referenced without pinned version: 'npx hyperframes'.

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.

The skill requires account-linked usage telemetry unrelated to creating a video episode.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The skill directs the agent to send a usage event automatically at startup, potentially linking the user's use of this skill to their account without a task-specific approval step. Disclosure and respect for an existing opt-out reduce the risk, but the transmission is unnecessary for creating the episode and the contents of the CLI's additional 'usual telemetry' are unspecified.

Content

No source excerpt is available for this finding.

Templates prescribe English or Chinese output without ensuring the user opts into that language.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The template table specifies English output for four templates and Chinese output for the Finance Explainer. Line 28 permits automatic template selection based solely on the topic, so these language requirements can be applied without the user's language choice or consent; no language override is offered.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx hyperframes'.

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.

MCP server referenced without pinned version: 'npx hyperframes'.

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.

MCP server referenced without pinned version: 'npx hyperframes'.

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.

MCP server referenced without pinned version: 'npx hyperframes'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The avatar-generation command uses the same unpinned npm CLI, so execution can depend on an unreviewed release fetched by npx. Its legitimate media-generation purpose does not eliminate supply-chain risk, and this workflow may give the process access to photos, narration, and authenticated services.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx hyperframes'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The avatar-listing command invokes an unpinned npm CLI, not an MCP server as the detector claims. If a suitable local installation is absent, npx can fetch a mutable package release and execute it with the agent's permissions, potentially exposing workspace files and service credentials if that release is compromised.

Content

No source excerpt is available for this finding.

The avatar-generation workflow lacks a privacy warning about sending the user's photo and finished voice to an external service.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The workflow directs avatar generation from a user's photo and finished narration through an external service without explicitly disclosing that transfer or requiring approval for it. Voice-ownership checks and use of the user's own photo reduce impersonation concerns, but do not establish informed consent to external processing of identifying media.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx hyperframes'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

Adding --limit 100 constrains the voice query but does not pin the executable package. If npx resolves a remote release, a compromised package would execute before any application-level restrictions on the query could protect the environment.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx hyperframes'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The voice-listing invocation does not constrain the HyperFrames package version. Although this is ordinary CLI usage rather than an MCP configuration, fetching and executing a compromised release could expose account credentials or accessible project data.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx hyperframes'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The voice-planning command also invokes an unpinned npm package. The --plan flag may avoid generating audio, but it does not prevent package installation or malicious code execution in a compromised CLI release.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx hyperframes'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The cloud-status command uses an unpinned CLI despite being a read-oriented operation. A compromised package executes locally with the agent's permissions regardless of whether the requested service operation only retrieves job status.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx hyperframes'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The preview command can resolve and execute a mutable npm release rather than a reviewed tool version. Its background mode does not cause the supply-chain weakness, but compromised code could remain active while accessing project assets.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx hyperframes'.

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The detached render launches the same unpinned npm CLI, creating another opportunity to execute an unreviewed package release. Background execution is legitimate for long renders, but a compromised executable could continue running with access to project files and inherited credentials.

Content

No source excerpt is available for this finding.