The Agent Skills Directory

[NO_CODE] (SAFE): The skill is entirely defined by Markdown instructions (SKILL.md, phases.md, output-template.md). No executable binary or script files (Python, JS, Shell) are included in the skill package, which significantly reduces the attack surface.- [DATA_EXFILTRATION] (SAFE): There are no commands that initiate network connections or attempt to read sensitive configuration files (e.g., credentials or environment variables). The output is limited to generating a Markdown document saved in the local user data directory.- [PROMPT_INJECTION] (SAFE): Analysis of the instructions reveals no patterns indicative of jailbreak attempts, safety bypasses, or instructions to ignore previous system constraints. The prompt logic is focused exclusively on a 3-phase design methodology.- [EXTERNAL_DOWNLOADS] (SAFE): The skill does not perform runtime downloads. While the documentation refers to a skill manager for installation, the operational logic of the skill itself is self-contained within the provided Markdown files.

project-designer