Back to skill

Security audit

Auto-AI Web Bridge

Security checks for vulnerabilities and agentic risk

Overview

This skill needs Review because it tells the agent to ask for and possibly save Claude/Gemini credentials while automatically sending user prompts or code to third-party AI websites.

Install only if you are comfortable with an agent using logged-in Claude and Gemini web sessions for your requests. Do not provide passwords, MFA codes, cookies, or recovery credentials to the agent; log in directly in the provider-controlled browser instead. Treat any code or prompt sent through this skill as shared with the external provider, and avoid using it for secrets, proprietary source, customer data, or regulated information unless you have explicit approval.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:41
Finding

Unsafe Collection and Ambiguous Storage of User Credentials

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
SKILL.md:9
Finding

Automatic Disclosure of User Prompts and Source Code to External AI Services

Content
View full analysis
Remediation
View remediation
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 (7)

Context-Inappropriate Capability

Critical
Category
Not specified by scanner
Confidence
99% confidence
Finding

The note to 'save credentials securely' confirms the skill contemplates persistent handling of user credentials, which is unrelated and excessive for a content-generation helper. Any persistence of third-party passwords materially increases the chance of credential theft, reuse compromise, and unauthorized account access.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The workflow sends user prompts, and for code tasks potentially proprietary or sensitive source code, to external services (Claude AI and Gemini) without explicit disclosure or consent. This is dangerous because it can leak confidential material to third parties under different retention, training, or access policies than the user expects.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill explicitly instructs the agent to ask for third-party account credentials to access Claude/Gemini, even though the skill’s stated purpose is only code or image generation. This creates unnecessary credential collection risk and could expose user passwords to the agent runtime, logs, or downstream components.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill tells the agent to request credentials for third-party sites but provides no explicit privacy or security warning about password handling. In context, this makes accidental disclosure more likely and normalizes unsafe credential-sharing behavior with the assistant.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The activation conditions cover broad, common requests such as any coding, code analysis, or image generation request, causing the skill to trigger frequently and potentially unexpectedly. In this context that is risky because activation leads to external website use, possible credential requests, and third-party data transmission without narrow scoping.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill instructs downloading generated files and saving them to a local path without clearly informing the user that a local file write will occur. Silent writes can create privacy, disk hygiene, and trust issues, especially if filenames or content are sensitive or later consumed by other tools.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

All user-facing instructions and activation text are presented in Italian, and the file does not indicate that this language restriction is optional or region-specific. This can amount to a language policy issue if the skill effectively forces one locale without user opt-in.

Content

No source excerpt is available for this finding.