保利威直播

• Outcome: Pass
• Latest audit: Jun 27, 2026 · 2:11 PM UTC
• Version: 1.0.9

Overview

This is a coherent PolyV live-stream administration skill, but it should only be used by authorized operators because it can change live resources and reveal viewer data.

Install only if you administer a PolyV account and trust the npm polyv-live-cli package. Before running mutating commands, confirm the active account, channel or resource ID, and whether the command affects global or production state. Treat exported JSON, viewer records, donation records, chat logs, check-in results, passwords, and AppSecret values as sensitive data.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (15)

Missing User Warnings

Medium

• Confidence: 85% confidence
• Finding: The documentation includes force-deletion examples such as `channel delete ... -f` and batch deletion without clear warnings about irreversibility or safeguards. In an agent skill context, this can normalize unsafe destructive operations and increase the chance that users or automation invoke irreversible channel deletion accidentally at scale.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The documentation exposes account-wide moderation actions such as global ban and global kick without prominently warning about their broad, potentially disruptive effect. An operator could unintentionally apply these actions beyond a single channel, causing denial of participation for many users and creating moderation abuse risk through simple misuse rather than exploitation of code.

Missing User Warnings

Low

• Confidence: 77% confidence
• Finding: The chat history command is documented as a routine listing operation without caution that outputs may contain message contents, nicknames, and other user identifiers. In a live-platform administration context, this can lead to unnecessary exposure of personal or sensitive communications when operators export or share results, especially in JSON output.

Missing User Warnings

Low

• Confidence: 81% confidence
• Finding: The documentation explicitly recommends exporting check-in results to a local JSON file, but gives no warning that the file may contain participant-level attendance data that could be retained, shared, or mishandled outside the platform. In a live-education or training context, this creates a real privacy and data-handling risk even though it is a documentation weakness rather than an exploit primitive.

Missing User Warnings

Low

• Confidence: 86% confidence
• Finding: The skill documents a command that returns detailed check-in results, including signed-in and unsigned users, without noting that this exposes participant-level attendance information. In this product context that capability is expected, but the absence of privacy guidance can still lead to overexposure or inappropriate sharing of sensitive participation data.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The command documentation explicitly exposes donation records containing viewer user IDs, nicknames, session IDs, timestamps, and amounts, but provides no privacy, authorization, retention, or handling guidance. This can encourage operators or downstream scripts to collect and export personally identifiable or behavior-linked data without adequate safeguards, increasing the risk of privacy violations or unauthorized profiling.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The documentation includes update and delete examples using --force, which bypasses confirmation prompts for actions that can modify or remove live lottery configurations. In an agent or scripted context, this increases the chance of accidental destructive changes during live operations because the examples normalize unsafe defaults without warning about operational impact.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The routing guide explicitly maps user requests to destructive channel operations, including deletion, without requiring a visible confirmation or warning at the routing stage. In an agentic setting, natural-language intent can be ambiguous, so this increases the risk of accidental or overly broad destructive actions against production live channels.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The skill routes requests directly to playback deletion operations without warning that removing playback may permanently affect audience access, records, or business evidence. Because this file is intended to drive command selection from free-form user language, omission of safeguards can lead to accidental destructive execution.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The file routes requests involving watch conditions, tokens, and whitelist management without any warning that these actions can change who is allowed to access a live stream. Misrouting or overbroad execution could unintentionally expose restricted content or lock out legitimate viewers, creating privacy and availability risks.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The routing document includes chat clearing, kicking, muting, and blacklist-related actions but does not disclose that some of these can remove participation history or disrupt user access. In a moderation workflow, these actions are operationally legitimate, but without confirmation and impact warnings they can be triggered too easily from ambiguous user requests.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The file documents passing `--appId` and `--appSecret` directly on the command line without warning that shell history, process listings, CI logs, and terminal recordings may expose secrets. In an admin-oriented skill managing live-service resources and viewer data, this increases the chance of credential leakage and subsequent unauthorized access.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The documentation explicitly describes commands that retrieve and list viewer personal data such as mobile numbers, email addresses, names, and areas, but it provides no privacy warning, access-control expectations, data-minimization guidance, or audit/consent considerations. In a live-stream administration CLI, this can normalize bulk access to PII and increase the risk of unauthorized collection, misuse, or accidental exposure by operators or downstream tooling.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The documentation includes examples that place watch passwords directly on the command line and in JSON files in plaintext. This is risky because command-line arguments may be exposed via shell history, process listings, logs, screenshots, and shared config files, leading to credential disclosure and unauthorized access to paid or restricted streams.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The documentation demonstrates collection of personal information such as name and mobile number for access control, but provides no privacy, consent, retention, or protection guidance. In a live-stream management context, this can lead operators to collect PII without adequate notice or safeguards, increasing the risk of privacy violations, regulatory noncompliance, and data exposure.

VirusTotal

65/65 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.