The Agent Skills Directory

[SAFE]: No malicious patterns such as prompt injection, data exfiltration, or obfuscation were detected. The instructions are purely functional and focused on guiding the user through an architectural design process.
[SAFE]: The skill does not perform any network operations or external downloads. All referenced materials are local markdown files included in the skill package.
[SAFE]: File system interactions are restricted to reading the project codebase for context and writing the final design document to a specific local directory (.gemini/technical-decisions/).
[NO_CODE]: This skill consists entirely of Markdown instructions and templates; it does not contain any executable scripts (Python, JavaScript, etc.), significantly reducing the attack surface.
[SAFE]: The operating mode explicitly forbids code implementation or silent assumptions while the skill is active, maintaining a clear separation between the design and implementation phases.

brainstorming