Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

auto-content-creator

v2.0.0

自动化内容创作助手 - 一键生成抖音/小红书/公众号内容。 支持:文生文案、图生文案、爆款改写、多平台适配。 触发词:"写文案"、"生成内容"、"小红书"、"抖音"、"公众号"、"content"、"create post"。 自动输出:标题 + 正文 + 标签 + 发布建议。

0· 73·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
Name, description, SKILL.md content and the included Python script all focus on multi-platform content generation (小红书/抖音/公众号/etc.). The functionality requested (reading topics or input markdown files, generating text) matches the stated purpose. Minor incoherence: SKILL.md uses a CLI name 'auto-content' in examples but there is no install spec or entrypoint that provides that command — only a scripts/content-generator.py file is present. That means the CLI examples are misleading unless the agent environment already has a wrapper or the user manually runs the script with python.
Instruction Scope
Runtime instructions and examples show the agent/user running the CLI and optionally providing input files (e.g., 原文.md, 话题列表.txt). Those file reads are reasonable for content generation. The SKILL.md does not instruct the agent to read unrelated system files, shell history, or environment secrets, nor to send data to external endpoints. However, batch and rewrite examples indicate the skill will process arbitrary user-supplied files, so users should avoid passing sensitive documents.
Install Mechanism
There is no install specification (low install-surface risk). The skill is 'instruction-only' in metadata but ships a Python script. Because no install steps create a CLI wrapper or place the script on PATH, the documented usage (auto-content create ...) is inconsistent with what's provided. This is not inherently malicious but is an incoherence: to use the script the environment must have Python and the operator must run the script directly or create their own wrapper.
Credentials
The skill declares no required environment variables, no credentials, and no config paths. The included script writes output under /tmp and otherwise appears to operate on supplied inputs; no network libraries (requests, urllib) are imported in the visible part, and no credential access is required. This is proportionate to a content generation tool.
Persistence & Privilege
Skill flags: always=false and normal autonomous invocation settings. The skill does not request persistent system-wide privileges, nor does it attempt to modify other skills or global agent configuration (based on the provided files).
What to consider before installing
What to consider before installing/using this skill: - The skill’s purpose and the included Python script align (content generation) and it asks for no secrets — that is good. - However, SKILL.md shows a CLI named 'auto-content' while the package only includes scripts/content-generator.py and no install instructions or entrypoint. Expect to run the script directly with python or create your own wrapper; the documentation is inconsistent. - The script writes output to /tmp and will read any input files you give it (batch, rewrite). Do not feed it sensitive documents or credentials. - The visible code does not import network libraries, which reduces exfiltration risk, but the provided file in this listing is truncated near save_result; the tail of the script could contain additional behavior (network calls, subprocess execution, telemetry). Because of that truncation, exercise caution: inspect the rest of scripts/content-generator.py (the full file) for any subprocess, open network sockets, HTTP/HTTPS requests, or code that reads files outside of user-supplied inputs before running it. - If you plan to run it, run inside a sandbox/container, or at least review the complete script and confirm there are no calls to subprocess, os.system, sockets, or external endpoints. Ask the publisher for an install script or a signed release, and prefer packages with a homepage or identifiable author if you need to use it in production.

Like a lobster shell, security has layers — review code before you run it.

latestvk9746wf2aem8ttad1f9q1jwq1n83m0sk

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments