The Agent Skills Directory

[SAFE]: The skill instructions define a comprehensive and security-conscious workflow for reviewing code changes. It explicitly cautions the agent against common pitfalls like trusting frontend-only controls or assuming parameterized SQL prevents all authorization issues.
[EXTERNAL_DOWNLOADS]: The documentation references installation and maintenance commands (npx skills add, npx skills update) targeting the author's GitHub repository (github.com/idaibin/aicraft). These are standard documented procedures for managing skills within this ecosystem and point to the vendor's primary infrastructure.
[COMMAND_EXECUTION]: The maintenance section mentions a local validation script (python3 scripts/validate-skills.py). This is presented as a manual step for skill developers rather than an automated or hidden execution pattern within the skill's runtime logic.
[PROMPT_INJECTION]: No evidence of prompt injection, jailbreak attempts, or system prompt extraction instructions was found. The instructions are focused on providing a professional security review service.

code-security