ClawHub
Back to skill

Security audit

Markdown转PDF（免费版）

Security checks across malware telemetry and agentic risk

Overview

This skill is a straightforward Markdown-to-PDF helper, with normal local file and command access for that purpose but some low-quality broad boilerplate users should read carefully.

Install only if you are comfortable with a local agent reading Markdown files, running pandoc/xelatex, and writing PDF outputs. Use explicit file paths, avoid sensitive directories, and confirm before overwriting existing files; the broad modify/delete wording appears to be boilerplate rather than a supported feature.

SkillSpector

Vulnerability Patterns
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The skill uses broad natural-language trigger phrases such as converting Markdown or batch-converting files, which can overlap with ordinary user conversation and cause unintended invocation. In an agent with read/write/exec permissions, accidental activation can lead to unreviewed file processing or command execution paths that the user did not explicitly authorize.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The operation scope is described in very broad terms, including create/query/export and even modify/delete style operation language, without clear boundaries on what files or actions are permitted. In a skill with exec, read, and write capabilities, this ambiguity increases the chance of overreach, unsafe tool use, or actions outside the user's intended conversion task.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill demonstrates creating an output PDF but does not warn about overwriting existing files or creating files in sensitive locations. In agent environments with write access, this can cause unintended data loss or unauthorized file placement if the output path is user-controlled or inferred incorrectly.

Unvalidated Output Injection

High
Category
Output Handling
Content
import subprocess
# ...
def md_to_pdf(md_file, output_pdf):
    result = subprocess.run([
        "pandoc", md_file,
        "-o", output_pdf,
        "--pdf-engine=xelatex",
Confidence
93% confidence
Finding
The example passes unvalidated input and output file paths directly into a subprocess invocation of pandoc. Even without shell interpolation, attacker-controlled paths or crafted documents can cause unsafe file reads/writes, overwrite arbitrary files accessible to the agent, or trigger risky behavior in downstream converters and LaTeX tooling during document processing.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.