Back to skill

Security audit

Employee Reminder Ops

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly aligned with employee reminder reporting, but its bundled scripts can post employee data to a real-looking default Discord channel with production-send automation.

Review this before installing in production. Replace all bundled spreadsheet, account, and Discord channel defaults with your own explicit configuration, use a test channel first, remove unnecessary Drive authorization, avoid broad .env sourcing where possible, and only enable scheduled prod-send after confirming the destination audience is approved for employee birthday and event data.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/plan-a-demo.js:8
Finding

Hardcoded production identifiers can route employee data to an unintended Discord channel

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
references/deployment.md:54
Finding

Deployment instructions request unnecessary Google Drive authorization

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
references/deployment.md:28
Finding

Documentation installs an unpinned global npm dependency

Content
View full analysis
Remediation
View remediation
`. 2. Document the expected npm registry, package publisher, and package provenance. 3. Verify package integrity or signed provenance where supported. 4. Prefer a project-local, lockfile-controlled installation over a global installation when practical. 5. Review package lifecycle scripts before recommending installation. 6. Use `npm ci` with a committed lockfile for reproducible dependency resolution where the deployment model permits it. 7. Establish an update process that reviews new versions before changing the pinned version. 8. Advise users not to run the installation command with elevated privileges. ]]>
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (19)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The declared description is specific: a Google Sheets-driven employee reminder and reporting workflow involving staff/event schemas, scheduled jobs, and chat routing. The provided code chunk does not implement or demonstrate any of that functionality directly. Instead, it is a thin launcher script that loads environment variables and runs an external Node script in 'prod-send' mode. Because the visible behavior is primarily orchestration of an unspecified production send operation, and there is no evidence in this chunk of Google Sheets access, reminder/report generation, or Telegram/Discord routing, the code does not accurately represent the declared purpose based on the supplied snippet.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · references/deployment.md (reported line 57)May include surrounding context.

Ví dụ flow:

bash
gog auth credentials /path/to/client_secret.json
gog auth add your-google-account@gmail.com --services sheets,drive

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding

The skill metadata does not declare any tool restrictions even though the workflow clearly involves environment access and outbound network actions to Google Sheets and chat platforms. Without explicit scope boundaries, an agent may invoke broader capabilities than intended, increasing the chance of secret exposure or unauthorized message sending during installation or execution.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The document provides commands that will send real messages to a live Discord channel using a bot token, but the warnings before those commands are weak and do not clearly foreground the external side effect at the point of execution. In an operational skill that reads real Google Sheets data and posts into team chats, this can cause unintended disclosure, spam, or accidental production actions during testing or migration.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The plan explicitly routes automated reports to a fixed Discord channel but does not mention any consent, audience restriction, or data-minimization control. Because the workflow processes employee birthdays and special events, this can expose internal personnel data to a broader chat audience than intended, especially if the channel membership is not tightly controlled.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The deployment guide instructs operators to configure Google auth and chat bot tokens but does not explicitly warn that these credentials must be kept out of committed files, chat transcripts, logs, and shared screenshots. In an ops-focused skill that routes messages to Telegram/Discord and uses Google access, omission of credential-handling guidance materially increases the chance of accidental secret exposure during setup or troubleshooting.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The document prescribes tab names and required column headers entirely in Vietnamese, including a suggested default tab name and all schema labels. Because no opt-in, alternative locale, or justification for a Vietnam-specific deployment context is provided, this appears to enforce a specific language/locale in natural-language instructions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This markdown file contains operational instructions in Vietnamese (for example, 'Hỗ trợ hiện tại' and 'macOS là môi trường hiện tại đã test Plan A nhiều nhất') while other parts are in English. That imposes a language expectation on users without any stated opt-in, language selection, or justification for a locale-specific audience.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · references/architecture.md (reported line 22)May include surrounding context.

md
## Scheduler
- use `launchd`
- example plist already bundled in scripts

## Notes
macOS là môi trường hiện tại đã test Plan A nhiều nhất.

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · references/macos.md (reported line 14)May include surrounding context.

md
## Scheduler
- use `launchd`
- example plist already bundled in scripts

## Notes
macOS là môi trường hiện tại đã test Plan A nhiều nhất.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The headings and instructions from L03 onward are entirely in Vietnamese, which imposes a language choice on users without opt-in. Under the policy, language-specific guidance should either offer a choice or clearly document why the locale restriction is necessary.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This markdown mixes English headings with core operational instructions in Vietnamese, such as setup and scheduler guidance. Because the file does not state that the skill is intended only for Vietnamese-speaking users or offer an alternate language, it may violate language/locale policy requiring user choice or clear justification.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

In send and prod-send modes, the script transmits employee birthdays, department information, event details, and optionally invalid-row details to a Discord channel without any interactive confirmation, recipient validation, or data-classification guardrail. In the context of an internal operations workflow, this can cause accidental disclosure of employee-related information to the wrong channel or broader audience, especially because channel IDs and tokens are environment-driven and the feature is designed for automated scheduled execution.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The script silently sources a local .env file with set -a, exporting all variables into the process environment without any notice or validation. In an agent-skill context, this is dangerous because it can implicitly pull in sensitive credentials or unsafe runtime settings from the local workstation, broadening what the downstream Node process can access and making behavior dependent on hidden local state.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The script directly invokes a production-send mode with no visible warning, dry-run default, or confirmation prompt. In a skill meant to handle reminders and chat routing, that makes accidental real-world delivery to production channels likely, which can leak employee data, spam internal communications, or trigger unintended operational actions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The natural-language instructions, headings, and operational guidance are presented in Vietnamese only. Under the language/locale policy, forcing a single language without opt-in or justification can be a policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

Natural-language policy requires flagging cases where a skill forces a specific language or locale without user opt-in. This file's operational instructions are entirely in Vietnamese and do not indicate that the skill is region-specific or that users may choose another language.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The natural-language instructions are presented almost entirely in Vietnamese, which effectively forces a specific language for operators reading the deployment guide. Under the stated policy, a fixed language is a concern unless the file offers user choice or clearly documents that the skill is intentionally region- or locale-specific.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

The manifest describes an internal reminder/reporting workflow around Google Sheets and chat routing, but this launcher script imports all variables from a local .env file before execution. Reading arbitrary local environment configuration is not described in the manifest and introduces a broader capability than the business-purpose description suggests.

Content

No source excerpt is available for this finding.