The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected in the skill instructions or research files. The content is educational and focuses on cognitive modeling.
[REMOTE_CODE_EXECUTION]: The installation instructions utilize standard platform package management commands (npx). No unauthorized or suspicious remote code execution vectors were found.
[COMMAND_EXECUTION]: No dangerous system commands, privilege escalation (sudo), or persistence mechanisms are present in the scripts or prompts.
[DATA_EXFILTRATION]: No patterns for secret harvesting or unauthorized data exfiltration were identified. The mentioned use of search tools is scoped to a research protocol for factual verification.
[PROMPT_INJECTION]: The role-play instructions provide character constraints and a specific 'Identity Card' but do not attempt to override system safety guidelines or extract system prompts.

feynman-perspective