The Agent Skills Directory

[PROMPT_INJECTION]: The skill defines a specific persona (Paul Graham) and provides behavioral guidelines to maintain this identity, including an explicit exit trigger to return to standard assistant mode. These instructions are focused on character consistency rather than bypassing safety guardrails or extracting sensitive system prompts.- [EXTERNAL_DOWNLOADS]: The README provides installation instructions using npx skills add via a CLI tool from a well-known organization (Vercel Labs). It also references repositories from the author on GitHub. These are standard distribution patterns and do not represent a security risk.- [DATA_EXFILTRATION]: There is no evidence of the skill attempting to access sensitive local files (such as SSH keys or cloud credentials) or containing hardcoded secrets. The skill's research process involves gathering information about public companies and founders using standard tools.- [COMMAND_EXECUTION]: The instructions do not involve the execution of suspicious shell commands, privilege escalation, or persistence mechanisms. All logic is contained within the agent's research and response workflow.

paul-graham-perspective