The Agent Skills Directory

[SAFE]: The skill consists solely of markdown documents providing guidelines, checklists, and theoretical frameworks. It does not include any scripts, binaries, or automated tools.\n- [PROMPT_INJECTION]: No prompt injection attempts were detected. The instructions do not attempt to override system safety guidelines, pretend to be an unrestricted AI, or ignore previous instructions.\n- [DATA_EXFILTRATION]: There is no evidence of network activity or access to sensitive local files (e.g., SSH keys, cloud credentials). The skill only interacts with the content of the business plan it is auditing.\n- [REMOTE_CODE_EXECUTION]: No remote code execution patterns or package installations were found. The skill does not download or execute scripts from external sources.\n- [COMMAND_EXECUTION]: The skill does not contain any shell commands or subprocess calls. There are no operations involving privilege escalation or persistence mechanisms.

meta-sustainability