Back to skill

Security audit

Turn photos, memories, and event details into polished custom invitation posters.

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Mew.design invitation-poster workflow, with expected use of user photos, event details, and a user-provided API key.

Install only if you are comfortable sending the selected photos, poster assets, event details, and your Mew.design API key to Mew.design. Avoid including private addresses, personal contact details, or sensitive images unless they are necessary for the invitation, and use the temporary third-party upload fallback only after considering the privacy tradeoff.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The install-intent examples are broad enough that an agent could match generic requests like 'help me install this skill' without a strong binding to this specific package. That increases the chance of unintended installation or invocation, especially in environments where multiple skills can respond to similar install language.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The embedded onboarding response is written entirely in Chinese and is presented as the copy to use when the user lacks an API key. The file does not indicate that Chinese should only be used when the user prefers it, creating a language/locale policy issue.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

Referencing the external domain itself is not the issue; the danger is that this endpoint receives personal content and authentication material. In the context of an image-processing skill, transmitting family, home, pet, or relationship photos to a third party can expose highly sensitive personal information if users do not fully understand the privacy implications.

Content

Scanner excerpt · SKILL.md (reported line 242)May include surrounding context.

For stage 1, call the official Mew.design image-process API directly:

bash
curl -sS -X POST "https://api.mew.design/open/api/image/process" \
  -H "Content-Type: application/json" \
  -H "x-api-key: USER_PROVIDED_KEY" \
  -d '{

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

Referencing the external domain itself is not the issue; the danger is that this endpoint receives personal content and authentication material. In the context of an image-processing skill, transmitting family, home, pet, or relationship photos to a third party can expose highly sensitive personal information if users do not fully understand the privacy implications.

Content

Scanner excerpt · SKILL.md (reported line 242)May include surrounding context.

For stage 1, call the official Mew.design image-process API directly:

bash
curl -sS -X POST "https://api.mew.design/open/api/image/process" \
  -H "Content-Type: application/json" \
  -H "x-api-key: USER_PROVIDED_KEY" \
  -d '{

External Transmission

Medium
Category
Data Exfiltration
Confidence
92% confidence
Finding

The second-stage design-generation call sends transformed images plus event copy such as names, venue, date, and RSVP details to an external service. This compounds privacy exposure because it can reveal both identity-related imagery and structured real-world event information that could be sensitive or exploitable if mishandled.

Content

Scanner excerpt · SKILL.md (reported line 256)May include surrounding context.

For stage 2, call the official Mew.design design-generate API directly:

bash
curl -sS -X POST "https://api.mew.design/open/api/design/generate" \
  -H "Content-Type: application/json" \
  -H "x-api-key: USER_PROVIDED_KEY" \
  -d '{

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The consent-first explanation for uploading a local image to a third-party host is hardcoded in Chinese. Since the file does not state that this locale is user-selected or region-specific, it may force a language choice on users who did not opt in.

Content

No source excerpt is available for this finding.