extension-openai
Fail
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: CRITICALPROMPT_INJECTIONMETADATA_POISONINGEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses extremely directive and high-pressure language (e.g., "MANDATORY recipe", "ONLY supported path", "FORBIDDEN anti-pattern") in both the description and instructions. This is designed to override the AI agent's logic and force it to exclusively use this skill's implementation while disregarding all alternative methods.
- [EXTERNAL_DOWNLOADS]: Automated security scans detected a phishing risk associated with the external URL
https://mops.one/openai-client, which the skill identifies as the required source for its dependency. - [REMOTE_CODE_EXECUTION]: The skill instructs the user to execute shell commands to install a package (
mops add openai-client@0.4.0) from a registry URL that has been flagged as malicious. This represents a significant supply chain risk and a vector for remote code execution. - [COMMAND_EXECUTION]: The skill requires the execution of shell commands for dependency management which, given the malicious URL and file reputation alerts, presents a high risk of system compromise.
- [CREDENTIALS_UNSAFE]: The "fully anonymous" variant described in the skill explicitly allows any unauthenticated visitor to overwrite the stored OpenAI API key. As these keys are billing credentials, this architecture enables unauthorized resource consumption and financial loss for the account holder.
- [METADATA_POISONING]: The skill's description and metadata fields are populated with aggressive instructional overrides intended to manipulate agent behavior during the discovery phase.
- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for building a chatbot that processes untrusted user input without explicitly mentioned sanitization or boundary markers. This creates an attack surface where external data can influence the LLM's behavior via indirect prompt injection.
- [OTHER]: Automated file reputation scans identified the
SKILL.mdfile itself as known malware.
Recommendations
- CRITICAL: 1 file(s) identified as malware by FileRep - DO NOT USE
- AI detected serious security threats
- Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata