The Agent Skills Directory

[SAFE]: The skill is purely instructional, providing system prompts and templates for security and architectural evaluation. No evidence of malicious obfuscation, unauthorized data access, or persistence was found.- [COMMAND_EXECUTION]: The skill uses allowed tools such as Read, Grep, and Bash to collect context (code or design decisions) from the environment for analysis. This behavior is consistent with the skill's stated purpose of auditing.- [PROMPT_INJECTION]: Persona definitions for security experts and architects are intended to provide high-quality technical feedback. They do not include instructions to bypass safety guardrails or disregard previous system constraints.

gpt-advisor