Back to skill

Security audit

md-web

Security checks for vulnerabilities and agentic risk

Overview

ClawScan's verdict is benign with medium confidence: it found no AIG issues and judged the reported behaviors generally consistent with the skill's purpose, but inspection was incomplete, so some risk remains uncertain.

The verdict is benign, but the evidence is incomplete: Skillspector reports only 46.7% coverage, with multiple files partially inspected, including `upload.js`. Consider this uncertainty before installing, especially since the tool scope is not declared; the verdict does not mean risk-free.

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
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (8)

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 48)May include surrounding context.

md
Exception: if it is missing but a legacy `config.json` sits next to `upload.js`, skip the wizard and just upload — the script migrates it (and `.deployed`) to `

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.

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.

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
80% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · README.md (reported line 26)May include surrounding context.

md
That pins the **installer**, not the **content**. The fetched skill is still whatever the
repository serves at install time; what pins the content is the skill's own released
version, so install a release you have read and re-run the install after an upgrade —
the version is in `skills/md-web/SKILL.md`. Reviewing the installer does not review the
skill: read the skill before you install it.

Or install manually — clone this repo and copy (or symlink) the `skills/md-web/` folder

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
80% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · README.zh.md (reported line 24)May include surrounding context.

md
固定的是**安装器**,不是**内容**。安装时取到的仍然是仓库当时的 skill;真正把内容钉住的是
skill 自己的发布版本 —— 请安装你读过的那个版本,升级之后再重装一次(版本号见
`skills/md-web/SKILL.md`)。审过安装器不等于审过 skill:装之前先读一遍。

或手动安装 —— clone 本仓库,把 `skills/md-web/` 目录拷贝(或软链)进你的 agent 读取的
skills 目录即可,只要是一个名为 `md-web` 的目录,具体路径按你的 agent 文档来。

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

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.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 104)May include surrounding context.

e two keys and exports MD_WEB_ACCESS_KEY / MD_WEB_SECRET_KEY instead (the file wins when both are set). 3. Optional: region (default auto; AWS needs the real region, e.g. us-east-1) and expire_days (auto-delete uploads after N days; 0 removes the rule). Omit expire_days unless the user asks for auto-expiry. Lifecycle administration needs provider-specific bucket permissions: R2 uses Admin Read & Write, while AWS uses lifecycle IAM permissions, not an R2 token label. The script only edits its own md-web/ rule and preserves other rules. 4. Write ~/.md-web/config.json (create the directory; on macOS/Linux create the directory with mode 0700 and the file with mode 0600 before writing secrets; the script also checks owner-only file access; it never asks for root and never needs elevated privileges):

json
{

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
70% confidence
Finding

Tool defaults are unsafe or overly permissive (e.g. disabled TLS verification, no authentication, world-writable permissions). Unsafe defaults widen the attack surface.

Content

Scanner excerpt · upload.js (reported line 281)May include surrounding context.

js
if (fs.existsSync(USER_CONFIG_PATH)) {
    CONFIG_PATH = USER_CONFIG_PATH;
  } else if (fs.existsSync(LOCAL_CONFIG_PATH)) {
    // Migrate credentials without a world-readable intermediate copy.
    try {
      migrateConfig(LOCAL_CONFIG_PATH, USER_CONFIG_PATH);
      CONFIG_PATH = USER_CONFIG_PATH;