offers
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill provides standard role-playing instructions for an expert consultant. No attempts to subvert agent safety protocols or bypass system constraints were found.
- [DATA_EXFILTRATION]: The skill does not contain any network-facing commands or logic to access sensitive system files. It refers to local project-specific marketing context files, which is a standard functional behavior for this use case.
- [REMOTE_CODE_EXECUTION]: There are no patterns involving the download or execution of remote scripts or the installation of third-party packages.
- [COMMAND_EXECUTION]: The skill consists entirely of Markdown text and does not invoke any shell commands or subprocesses.
- [OBFUSCATION]: No encoded content, zero-width characters, or other methods of concealing malicious text were identified.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user input regarding business offers but does not possess any dangerous capabilities (e.g., file writing, network calls) that would make it vulnerable to indirect injection attacks.
Audit Metadata