Back to skill

Security audit

Find Skills.Bak

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it says, but it encourages broad activation plus unpinned global skill installation with confirmations skipped, which can permanently change the user's agent environment.

Install only if you specifically want an agent helper that searches for and installs other skills. Avoid letting it run global or unattended installs by default; review the exact skill source, prefer pinned versions or trusted repositories, and require confirmation before adding or updating skills.

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

Warning
Location
SKILL.md:23
Finding

Unpinned Third-Party CLI Execution and Unattended Global Skill Installation

Content
View full analysis
` - Install a skill from GitHub or other sources - `npx skills check` - Check for skill updates - `npx skills update` - Update all installed skills ``` ```markdown ### Step 4: Offer to Install If the user wants to proceed, you can install the skill for them: ```bash npx skills add -g -y ``` ``` ### Technical Analysis The skill instructs the agent to execute `npx skills` without pinning the npm package to a reviewed version or verifying its integrity. Depending on the local npm and `npx` configuration, this can download and execute the package currently resolved by the configured registry. Its effective code may therefore differ from the version that was previously reviewed. The CLI is then used to install skills from GitHub or other sources. The documented installation command combines global installation (`-g`) with unattended confirmation (`-y`). The workflow does not require validation of the source owner, review of downloaded files, immutable commit pinning, checksum verification, package provenance verification, or final informed approval before installation. This creates a supply-chain trust boundary in which compromised, replaced, typosquatted, or otherwise malicious CLI packages and skills could execute under the privileges of the agent process. ### Attack Path 1. An attacker compromises, replaces, or impersonates the npm package resolved as `skills`, or publishes a malicious skill through a source accepted by the CLI. 2. The ...[truncated 1424 chars]
Remediation
View remediation
find ``` 2. Configure npm to use a trusted registry and verify package provenance, signatures, and integrity metadata before execution. 3. Prefer installing the CLI as a locked project dependency rather than downloading and executing the currently published package through an unpinned `npx` command. 4. Restrict skill sources to an explicit allowlist of reviewed owners and repositories. 5. Pin skill installations to immutable commit hashes or verified release tags rather than mutable branches or unversioned identifiers. 6. Download and inspect skill contents before activation, including instructions, scripts, package manifests, lifecycle hooks, and external network operations. 7. Remove `-y` and require explicit, informed user confirmation after presenting the exact source, revision, requested installation scope, and review results. 8. Prefer project-local or sandboxed installation instead of `-g` to limit the affected scope. 9. Run installation in a least-privileged environment without unnecessary credentials or access to sensitive files. 10. Do not perform bulk updates without reviewing release changes and revalidating the integrity and provenance of every updated component. ]]>
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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (17)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The description says the skill should be used whenever users ask broad questions like 'how do I do X' or express interest in capabilities. Such broad triggers can cause the skill to activate in many normal conversations and steer users toward package discovery/installation when they did not explicitly request code execution or environment changes.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The 'When to Use' section contains expansive, ambiguous conditions like any specialized capability request or general interest in extending capabilities. Over-broad activation increases the chance of unnecessary tool/package recommendations and can socially engineer users into installing third-party code as part of ordinary help-seeking.

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 skills without pinning a specific package version, which causes execution of whatever version is currently resolved from the registry. That creates a supply-chain risk: a compromised, typosquatted, or newly malicious release could be fetched and executed at runtime in the user's environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

This command reference uses npx skills without a fixed version, allowing arbitrary future package contents to be executed from the package registry. In a skill whose purpose is to discover and install additional code, unpinned execution materially increases supply-chain exposure.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The unpinned npx skills add <package> example can fetch and execute an unspecified version of the CLI, making the exact behavior dependent on registry state at execution time. That makes compromise of the CLI package or its dependency chain directly exploitable.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx skills check without pinning still executes remote package code whose version may change unexpectedly. Even seemingly read-only operations are dangerous because package install hooks or CLI startup code can run before any benign command logic.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

An unpinned npx skills update reference compounds risk because it both executes an unpinned CLI and performs update actions on installed skills. An attacker controlling the resolved CLI version could influence update behavior or install malicious code.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The search command example again relies on unpinned npx skills, exposing users to registry-time package substitution. Since this is presented as routine workflow, it normalizes repeated remote code execution without version control.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The example output instructs installation using npx skills add <owner/repo@skill> without pinning the CLI version. This is dangerous because it encourages users to execute mutable registry content as part of installing further third-party code.

Content

No source excerpt is available for this finding.

Rp1

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.

Rp1

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.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The installation example tells users to run npx skills add ... without pinning the CLI version. That exposes users to mutable package resolution at the exact moment they are about to install additional external code, increasing blast radius.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
97% confidence
Finding

The instruction to install with npx skills add <owner/repo@skill> -g -y is especially risky because it combines unpinned CLI execution with global installation and suppressed confirmations. This reduces user visibility while modifying the environment, making supply-chain compromise more severe.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill recommends npx skills add <owner/repo@skill> -g -y, which performs global installation and suppresses confirmation prompts without warning about trust, persistence, or system modification. This reduces friction around installing third-party code and can lead to silent, lasting compromise of the user environment if a malicious or vulnerable skill is installed.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The example repeats unpinned npx skills usage in a context that teaches installation behavior. Repetition increases the likelihood that users and downstream agents adopt unsafe patterns for executing registry-fetched code.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

Suggesting npx skills init without pinning still invokes mutable remote package code. Although initialization seems low risk, the danger occurs before command semantics matter because package code must first be resolved and executed.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

This additional unpinned npx skills reference continues the same supply-chain vulnerability pattern. In aggregate, the skill normalizes executing an unversioned package manager entrypoint across many workflows, magnifying exposure.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.