Back to skill

Security audit

AutoSkills CLI

Security checks for vulnerabilities and agentic risk

Overview

This skill is an auto-installer that may change a user's agent setup, and its instructions run an unpinned remote npm package without clear confirmation or containment guidance.

Review this carefully before installing. Only run it if you trust the npm package publisher and version, prefer a pinned exact version in an isolated environment, and use detection/listing modes before allowing it to install skills or activate integrations. Avoid running the global install unless you intend to keep the tool available across sessions.

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:31
Finding

Execution of Unpinned Third-Party npm Package

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:31-32, SKILL.md:41-46, SKILL.md:63-66
Vulnerability Type: Unpinned and unaudited third-party package execution
Risk Level: Medium

Vulnerable Code Snippets

bash
npx our-autoskills
bash
npx our-autoskills
npx our-autoskills detect
npx our-autoskills install
npx our-autoskills list
npx our-autoskills doctor
bash
npm install -g our-autoskills
npx our-autoskills

Technical Analysis

The documentation instructs users to execute our-autoskills directly from the npm registry without pinning an exact version or verifying an integrity digest. Depending on the local npm configuration and cache state, npx our-autoskills can retrieve the current registry release and immediately execute its CLI code.

The project contains only SKILL.md; it does not include the package implementation, a lockfile, dependency metadata, or integrity information that would allow the executed code and its transitive dependencies to be reviewed as part of this audit. The documented global installation also creates a longer-lived installation in the user's npm environment.

The package is described as installing additional Agent Skills. Consequently, compromise of the npm package, its maintainer account, or one of its transitive dependencies could introduce additional unreviewed components and expand the supply-chain exposure.

No evidence establishes that the current npm package is malicious. The risk arises from executing mutable, externally hosted code without version pinning or integrity verification.

Attack Path

  1. An attacker compromises the npm package, a maintainer account, or a transitive dependency, or publishes a malicious future release.
  2. A user follows the instructions in SKILL.md and runs npx our-autoskills or installs the package globally.
  3. npm resolves the mutable package release from the registry.
  4. Package li ...[truncated 905 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin the package to an exact reviewed version rather than resolving the latest release:
    bash
    npx --yes our-autoskills@<reviewed-exact-version>
    
  2. Record and verify the package tarball integrity digest before execution.
  3. Include the auditable implementation in the project or reference a reproducible, immutable source revision.
  4. Add a lockfile and review all direct and transitive dependencies.
  5. Avoid global installation unless it is strictly necessary; prefer an isolated, least-privileged environment.
  6. Disable npm lifecycle scripts during installation where functionality permits, and separately review any scripts that must run.
  7. Require explicit user confirmation before installing each additional Skill.
  8. Pin every installed Skill to an immutable reviewed version and verify its provenance and integrity.
  9. Run installation and detection operations without elevated privileges and with minimal access to secrets and sensitive directories.
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
92% confidence
Finding

The trigger phrases autoskills, 项目类型检测, and 智能安装 are broad and likely to appear in ordinary discussion, making unintended invocation more plausible. Because the skill can recommend and install components automatically, accidental triggering is more dangerous than for a purely informational skill.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill markets automatic detection and installation of skills but does not warn users that it may modify their environment, install components, or activate integrations. Lack of disclosure reduces informed consent and increases the chance of users triggering system-changing actions in contexts where only analysis or recommendation was expected.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The skill instructs users to run npx our-autoskills without pinning an exact package version, which causes execution of whatever package version is current in the registry at invocation time. Because this skill is explicitly about automatic detection and installation of other skills, an upstream package compromise or malicious update could immediately lead to arbitrary code execution and further system modifications.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

This unpinned npx command executes a package fetched at runtime, allowing registry drift or package takeover to change what code runs after the skill is published. In the context of an auto-installer, that risk is amplified because the executed package may then detect projects, install additional components, and modify the local environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx our-autoskills install without version pinning permits execution of whichever registry version resolves at the time, creating a supply-chain execution risk. Since this subcommand is specifically for installation, a malicious or compromised package could silently install unwanted or dangerous components.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding

Although list is less obviously dangerous than install, the unpinned npx invocation still executes remote package code before listing anything. A compromised package could abuse the trust implied by this harmless-looking command to run arbitrary code or stage follow-on actions.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

The doctor command appears diagnostic, but invoking it through unpinned npx still trusts the latest remotely served package. Because diagnostics often inspect system state, a malicious package could harvest environment details, enumerate installed tools, or prepare targeted abuse.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding

Another unpinned npx reference increases the chance that users copy-paste a runtime-fetched command whose behavior can change over time. Repetition across the documentation normalizes unsafe execution and broadens exposure to package-replacement or compromised-release scenarios.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The system-check section again directs users to execute an unpinned remote package. Given this skill's purpose of discovering environment details and managing skill installation, a compromised package could combine reconnaissance with unauthorized modifications or persistence.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The installation instructions recommend npx our-autoskills as a primary path, which fetches and executes the package without a fixed version and without any trust-verification mechanism. In a package meant to automate further installations, this creates a strong supply-chain risk with potentially cascading system impact.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

The manifest description and user-facing markdown are written entirely in Chinese, with no indication that the skill supports other languages or that Chinese is a required locale. This can violate language/locale policy when the skill is presented to a broader user base without opt-in.

Content

No source excerpt is available for this finding.