Back to skill

Security audit

装前先给一枚 skill 体检·逐条挂原文

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently drafts source-linked reviews without running the reviewed code, with optional external parsing that users control.

Use local-only review for confidential material. Approve external parsing only for URLs or documents you intend to share, excluding embedded tokens or signed links. Prefer a reviewed installer version and immutable source revision, or copy the inspected files manually. The checker validates report structure, not whether another skill is safe; external service behavior was not verified.

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 Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (6)

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 75)May include surrounding context.

md
bash scripts/fixtures/run_fixtures.sh # 离线三档一次跑完,末行自报档数与不合枚数

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

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

The installation command invokes an unpinned npm CLI through npx and selects no immutable revision of the skill repository. Although this is a CLI rather than an MCP server, compromised upstream releases could execute code with the installing user's privileges; the documented read-only review workflow does not protect this installation step.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

The recommended installation command leaves both the npx CLI version and the skill repository revision unpinned. This is not an MCP server reference, but it creates a real supply-chain exposure: a compromised CLI release could execute with the user's privileges, and mutable skill content could change after review.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_read, file_write.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The remote-review workflow lacks a privacy warning about submitting target URLs to an external parsing service.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The remote-review route sends target URLs to an external parsing service without warning that URLs can contain private repository identifiers, access tokens, or signed query parameters. Asking before spending does not establish informed consent to disclose that information, and testing the Bridge with a real local file also needs an explicit data-handling check.

Content

No source excerpt is available for this finding.

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

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.