sayhello
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- [PROMPT_INJECTION] (SAFE): No patterns of instruction override, jailbreak attempts, or system prompt extraction were found in the skill's markdown or script logic.
- [DATA_EXFILTRATION] (SAFE): The skill does not perform network operations or access sensitive file paths (e.g., credentials, SSH keys). It operates entirely on local, hardcoded data.
- [REMOTE_CODE_EXECUTION] (SAFE): No remote code execution patterns, such as piped curl-to-bash or dynamic loading from untrusted URLs, were detected.
- [COMMAND_EXECUTION] (SAFE): The provided Python script uses the standard library to process arguments and format strings. It does not spawn subprocesses or execute shell commands.
- [INDIRECT_PROMPT_INJECTION] (LOW): The skill processes user-supplied parameters for language, style, and context.
- Ingestion points: Command-line arguments in
scripts/greet.py. - Boundary markers: Not present.
- Capability inventory: Limited to printing hardcoded greeting strings to stdout.
- Sanitization: Input values are validated against keys in a hardcoded dictionary before being used to select a greeting. The risk is negligible as the output cannot be influenced to contain malicious instructions beyond the predefined list.
Audit Metadata