Security audit

邮件摘要免费版

Security checks for vulnerabilities and agentic risk

Overview

This skill is a mailbox-summary helper that openly uses a local browser session to read inbox metadata and save a local report, but users should treat its outputs as sensitive.

Install only if you are comfortable letting the agent use your logged-in browser to view your mailbox. Run it only for mailboxes you own or are authorized to access, and store or delete the generated screenshots and Markdown reports carefully because they may include private senders, subjects, snippets, and timestamps.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Vague Triggers

Medium

Confidence: 82% confidence
Finding: The skill description broadly advertises automated login and mailbox scraping without clearly constraining when the agent may access a live mailbox or what explicit user consent is required. In an agent setting, ambiguous activation language increases the chance of unintended execution against a real authenticated session, causing privacy-invasive access to sensitive email data.

Missing User Warnings

High

Confidence: 95% confidence
Finding: The skill omits a clear warning that it will access live mailbox contents and create local reports and screenshots containing potentially sensitive information. Without explicit disclosure, users may invoke it without understanding that confidential senders, subjects, snippets, and visual inbox captures will be stored or exposed locally.

Ssd 3

High

Confidence: 97% confidence
Finding: The skill explicitly instructs logging into webmail, extracting message metadata and snippets, and producing summaries from a live mailbox. This creates a direct path for collection and disclosure of sensitive communications data, especially when reused authenticated browser sessions are involved.

Ssd 3

High

Confidence: 98% confidence
Finding: The core capability section describes extracting inbox records and archiving screenshots, which can preserve sensitive message details far beyond the immediate task. Screenshots and structured summaries materially increase the risk of secondary disclosure through local files, logs, sync folders, or later agent access.

Ssd 3

High

Confidence: 97% confidence
Finding: The workflow operationalizes extraction of mailbox records and visual copies of the inbox, turning sensitive-access behavior into a step-by-step execution path. Because it uses a real browser session, the agent could access actual authenticated mailboxes and exfiltrate or persist private content with little friction.

Ssd 3

High

Confidence: 96% confidence
Finding: The example output demonstrates disclosure of sender identities, subjects, timestamps, and counts in a generated report, normalizing exposure of mailbox content in plaintext. Such examples encourage implementations that persist or display private communication details without redaction or access controls.

Static analysis

No suspicious patterns detected.