Xiaohongshu Assistant

Security checks across malware telemetry and agentic risk

Overview

The visible assistant mostly generates Xiaohongshu copy, but the package also includes under-declared code for logging in, publishing, commenting, and liking despite claiming those features were removed.

Install only if you intend to use it as a copywriting assistant and are comfortable removing or ignoring the bundled publisher/account modules. Do not provide Xiaohongshu login/session access or run xhs_publisher.py unless you have reviewed the external XiaohongshuSkills scripts and require explicit confirmation before any publish, comment, like, or bookmark action.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

#
ASI02: Tool Misuse and Exploitation
Medium
What this means

If this code is invoked, it could post to a user's Xiaohongshu account without the manual review workflow the README recommends.

Why it was flagged

This module can perform an actual headless publish to Xiaohongshu rather than only generate copy; the README says automatic publishing is removed/disabled.

Skill content
def publish(title, content, images=None, account="default", preview=False): ... "--title", title, "--content", content ... else: cmd.append("--headless")
Recommendation

Remove the publisher module from the skill package or clearly document it as an optional high-impact capability with explicit user confirmation and preview-only defaults.

#
ASI03: Identity and Privilege Abuse
Medium
What this means

The skill package may use or affect logged-in Xiaohongshu account sessions if the publisher/login helpers are called.

Why it was flagged

The code includes account login/session handling paths, while the registry declares no credentials or required config and the README says account automation is disabled.

Skill content
def login(account="default"): ... cmd = ["python", str(LOGIN_SCRIPT), "login", "--account", account]
Recommendation

Declare any account/session requirements in metadata, scope them narrowly, and require explicit user approval before any login or account action.

#
ASI04: Agentic Supply Chain Vulnerabilities
Medium
What this means

If a sibling XiaohongshuSkills directory exists, this skill could run unreviewed external automation code with the user's account context.

Why it was flagged

The publisher depends on scripts outside the provided package, so the code that would handle login and publishing is not reviewed, pinned, or declared by the install spec.

Skill content
XHS_SKILLS_DIR = Path(__file__).parent.parent.parent / "XiaohongshuSkills"; PUBLISH_SCRIPT = XHS_SKILLS_DIR / "scripts" / "publish_pipeline.py"; LOGIN_SCRIPT = XHS_SKILLS_DIR / "scripts" / "cdp_publish.py"
Recommendation

Vendor or pin the helper code, include it in the reviewed manifest, or remove the external execution path entirely.

#
ASI09: Human-Agent Trust Exploitation
Medium
What this means

Users may believe the skill cannot affect their Xiaohongshu account when the package still contains code capable of account automation if invoked.

Why it was flagged

This safety claim conflicts with bundled xhs_publisher.py code that still exposes publishing, login, commenting, and interaction wrappers.

Skill content
⚠️ **自动发布功能已移除**(法律和安全风险) ... ✅ 仅文案生成(零风险)
Recommendation

Avoid absolute claims like “zero risk,” and align the README, metadata, and package contents by removing or fully disclosing disabled high-impact code.