proof
Pass
Audited by Gen Agent Trust Hub on Feb 28, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill provides instructions for a specific text-processing task (proofreading) and does not exhibit any malicious behaviors.- [NO_CODE]: The skill is composed strictly of Markdown-based instructions and metadata. It does not include Python scripts, JavaScript, shell commands, or any other executable logic.- [DATA_EXPOSURE]: There is no evidence of the skill attempting to access sensitive files (e.g., SSH keys, AWS credentials) or environment variables.- [REMOTE_CODE_EXECUTION]: No patterns for downloading or executing remote code (such as curl/wget piped to a shell) were found.- [PROMPT_INJECTION]: The skill does not contain instructions that attempt to bypass safety filters or override the agent's core system prompts. It uses standard instructional language to define its scope.- [INDIRECT_PROMPT_INJECTION]: While the skill is designed to process untrusted user input (text for proofreading), it lacks any dangerous capabilities (like file writing or network access) that could be exploited via an indirect injection attack. It acts solely as a text transformer.
Audit Metadata