Back to skill

Security audit

邮件日程

Security checks for vulnerabilities and agentic risk

Overview

This skill reads local Apple Mail data to create local reminders, which is sensitive but clearly tied to its stated purpose and not hidden or deceptive.

Install only if you are comfortable letting the skill read Apple Mail sender, subject, and summary data for the selected range and create reminders in your default Reminders list. Review the remindctl Homebrew tap before installing it, and run narrow ranges such as today or unread instead of all unless needed.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:77
Finding

Unpinned Executable Installed from a Third-Party Homebrew Tap

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 77-80
Vulnerability Type: Unpinned third-party executable dependency
Risk Level: Medium

Vulnerable Code

markdown
## 依赖

- macOS 自带邮件应用
- Python 3
- sqlite3
- remindctl (`brew install steipete/tap/remindctl`)

Technical Analysis

The installation instructions direct users to install remindctl from the third-party Homebrew tap steipete/tap without pinning an immutable version, commit, package digest, or checksum. Consequently, the executable installed by this command can change after the Skill has been reviewed.

The installed executable is later resolved through the process PATH and invoked by scripts/create_reminders.py. Email-derived titles, sender addresses, summaries, and due dates are passed to that executable as command-line arguments. Although the Python code safely uses an argument array rather than shell=True, this does not protect against a compromised dependency or a malicious executable placed earlier in PATH.

There is no evidence in the audited project that the referenced package is currently malicious. The vulnerability is the absence of dependency pinning and integrity verification for a third-party executable that processes sensitive data.

Attack Path

  1. An attacker compromises the third-party Homebrew tap, its source repository, release artifacts, or distribution credentials.
  2. The attacker publishes a modified remindctl package under the same package name.
  3. A user follows the documented command:
    bash
    brew install steipete/tap/remindctl
    
  4. Homebrew installs the modified executable because the instructions specify no audited version or artifact checksum.
  5. The user runs the documented email-to-reminder workflow.
  6. create_reminders.py resolves and executes remindctl through PATH.
  7. The compromised executable receives email-derived data and executes with the permissions of the user running the Skill.

A related l ...[truncated 857 chars]

Remediation
View remediation

Remediation Suggestions

  1. Pin remindctl to an audited, immutable release or source commit rather than installing the mutable head of a third-party tap.
  2. Publish the expected SHA-256 checksum or other cryptographic digest for the approved artifact and require verification before installation.
  3. Prefer a trusted, reproducible package source with signed releases and documented provenance.
  4. Record the reviewed dependency version in SKILL.md and provide an explicit upgrade and re-audit process.
  5. Resolve the approved executable during setup and invoke it through a verified absolute path rather than relying solely on PATH.
  6. Before execution, verify that the executable path, ownership, permissions, version, and checksum match the approved installation.
  7. Minimize sensitive data passed to the dependency. Avoid including email summaries or sender information unless required for reminder creation.
  8. Document that users should review the dependency's source and permissions before granting terminal or automation access to Mail and Reminders.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (9)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly instructs reading from the local macOS Mail database and processing message subjects/content, but it does not prominently warn the user that potentially sensitive email content will be accessed. This creates a privacy and consent risk because users may trigger the skill without understanding the scope of local data exposure, especially since email often contains confidential personal or business information.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill automatically pipes retrieved emails into a reminder-creation script, but the description does not clearly warn that reminders will be created in the user's default reminders list. This can cause unintended state changes, privacy leakage via reminder titles/times derived from email content, and user confusion if reminders are created without explicit informed consent.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This script automatically creates reminders from untrusted email content without any confirmation, approval step, or sender trust check. An attacker who can send crafted emails could cause unwanted reminder creation, leading to spam, social-engineering amplification, or nuisance persistence in the user's reminder system.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/create_reminders.py (reported line 114)May include surrounding context.

python
'--notes', notes
    ]
    
    result = subprocess.run(cmd, capture_output=True, text=True)
    return result.returncode == 0

def main():

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The script directly accesses the user's Apple Mail database and exports message metadata and summaries without any explicit consent prompt, privacy disclosure, or access restriction. In an agent-skill context, this is dangerous because it enables bulk extraction of sensitive personal communications data, which could be exposed to other components, logs, or remote services.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

Comments, usage text, parsing expectations, and user-visible output strings are written in Chinese, which imposes a single language/locale on users. The file does not offer any language choice, opt-in, or documented justification for restricting operation to a Chinese-language context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The script's human-facing comments and error/usage strings are written only in Chinese, which imposes a specific language without any opt-in or justification. Under the language/locale policy, this is a natural-language constraint that should either be optional or explicitly documented as region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

The missing-database error is presented only in Chinese, which can prevent comprehension for users who did not opt into that locale. This is a natural-language policy issue because the script fixes the interaction language without presenting a choice or documented reason.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

The usage/error output for invalid range values is only in Chinese, again enforcing a specific language for user interaction. Without user opt-in or justification, this conflicts with the locale-choice policy.

Content

No source excerpt is available for this finding.