cowork

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [PROMPT_INJECTION] (SAFE): The instructions consist of behavioral guidelines and professional etiquette rules. There are no attempts to bypass safety filters, override system prompts, or extract internal instructions.
  • [DATA_EXFILTRATION] (SAFE): The skill does not contain network request code (curl, wget) or references to sensitive system files (SSH keys, AWS credentials).
  • [COMMAND_EXECUTION] (SAFE): No shell commands or executable scripts are present in the skill.
  • [INDIRECT_PROMPT_INJECTION] (SAFE): The skill defines patterns for interacting with local directories like vault/ and planning/. However, it does not include any tools or code that could be used to exploit malicious content found in those files, such as network exfiltration or command execution capabilities.
  • [REMOTE_CODE_EXECUTION] (SAFE): No external dependencies or remote script downloading patterns were identified.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:20 PM