Back to skill

Security audit

Ultravibe Review

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious with medium confidence because its review instructions reportedly conflict with read-only assurances and repeatedly use package commands that may fetch unpinned code.

Consider the reported write requirements and package-download risk before installing. The scan covered one Markdown file but was marked partial because referenced files were missing; it does not confirm what those files or the workflow do, so the evidence is incomplete.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (14)

The skill promises a fully read-only review but instructs agents to create report files without disclosing an exception.

Medium
Category
Output Handling
Confidence
98% confidence
Finding

The read-only assurance conflicts with mandatory report generation and checkpoint advancement elsewhere in the skill. This overlaps with SDI-4 rather than demonstrating a separate exploit, but it can mislead users about authorized side effects and encourage agents to disregard the stated boundary.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
87% confidence
Finding

This is not an MCP server reference, but it presents a conditional package-execution risk: if a local compiler is unavailable, npx tsc can resolve the registry package named tsc, rather than the intended TypeScript compiler. The --noEmit option does not prevent package acquisition or malicious executable behavior, undermining the read-only assurance.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

This is a package-execution issue, not an MCP configuration issue: npx tsx can acquire and execute an unpinned registry package if no local installation is available. Because this command is the mandatory preflight step, an unsafe dependency can execute before any proposed checks provide protection; a verified, locked local installation would reduce the risk.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

The checkpoint-read command uses the same potentially unpinned npx tsx execution path, which can fetch executable code when the local dependency is missing. Naming the operation 'read' does not constrain the runner's permissions or prevent package-management side effects.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

The timeout check repeats the conditional supply-chain exposure from invoking npx tsx without ensuring a pinned local dependency. A compromised runner executes before it can enforce the intended timeout protection, although no compromised package is demonstrated here.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

The phase-entry command can execute a registry-resolved tsx version if a suitable local dependency is absent. This weakens the entry gate because the gate itself runs through an executable whose provenance is not established by the supplied file.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

Architecture-phase dispatch repeats the potentially unpinned npx tsx execution path rather than referencing an MCP server. If registry fallback occurs, an unreviewed runner executes with the agent's permissions before generating the subagent brief.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

The architecture-phase transition uses npx tsx without establishing a pinned local runner, creating conditional supply-chain exposure. A compromised runner could undermine checkpoint or transition validation; the supplied file does not establish that such compromise exists.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

Defect-scan dispatch repeats the same conditional unpinned-package execution risk through npx tsx. Mandatory delegation does not mitigate a runner compromise, because executable code runs in the parent context before the brief is handed to a subagent.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

The defect-scan transition again permits registry resolution of tsx if the dependency is absent locally. This is a repeated instance of the same supply-chain weakness, and the transition's validation purpose does not make the runner trustworthy.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

Security-review dispatch relies on the same potentially unpinned npx tsx runner. Running security checks later does not protect against malicious code executed while acquiring or starting that runner.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

The security-phase transition repeats the conditional package-execution exposure associated with npx tsx. A gate that validates security events cannot establish trust in the executable already running it, though a verified locked local dependency would substantially reduce this concern.

Content

No source excerpt is available for this finding.

The workflow requires creating report files despite explicitly declaring that all operations are read-only.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The workflow explicitly requires writing .reports/review-report.json despite an unconditional prohibition on creating or modifying files. This is an authorization-boundary inconsistency that could cause unauthorized report creation or overwriting, although the specified output paths limit the demonstrated impact.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
83% confidence
Finding

The final report transition invokes the same potentially registry-resolved tsx executable. This repeats the supply-chain risk at the completion boundary, where a compromised runner could affect report integrity or workflow state.

Content

No source excerpt is available for this finding.